Thesis · about 4 min read

The thesis

The problem was never collecting the data.
It's that nobody taught the machine what any of it means.

Every venue, festival and activation in the country is already generating more data than its operator can read. Ticketing, access control, Wi-Fi, EFTPOS, cameras, staffing rosters, bar tills, rideshare drop-offs. The gap is not another sensor and it is not another dashboard. The gap is context — the human logic that says which of those numbers matter, how they connect, what a good number looks like for this site, on this day, in this weather, with this crowd.

That judgement currently lives in the heads of the people who run the event. It gets applied once, verbally, in a debrief — and then it evaporates. The next event starts from zero.

CrowdSense is the apparatus for not starting from zero. It is powered by Artefact Group because Artefact is already in the room: building the site, running the show, hiring the crew, hanging the screens. Every event we deliver is a labelled training set produced as a by-product of work a client is already paying us for. The data acquisition cost is negative — we are paid to generate it.

The compounding claim, stated plainly

More events → more context → a sharper collective model → better decisions sold back into the next event. The same way a studio gets better by bringing on more freelancers and contractors: the network's judgement accretes even as individuals rotate through. Ours accretes into a system that does not leave when the contract ends.

Won contract value FY26

$527k

up 52% on FY25 · 31 deals

3-year growth

100%

CAGR FY23→FY26 on won value

Events delivered

150

83 InStitchu pop-ups alone · not signed deals

Live pipeline

$8.5k

the number this strategy exists to fix

The turnaround, stated precisely — Xero, 5 August 2026

H1 FY2026 result

−$36,150

Jul–Dec 2025 · income $168,004

H2 FY2026 result

+$48,027

Jan–Jun 2026 · income $277,998

Half-on-half swing

$84,177

the year turned, and stayed turned

Working capital

−$13,491

was −$26,511 · moves with AR/AP timing week to week

The honest read on the current year. FY2027 is five weeks old. Income is $46,667 — against roughly $48,610 in the equivalent window last year, so revenue is tracking marginally below last year, not above it. Reported net profit for the window is $32,298, but that flatters twice over: it includes a $12,001 ATO clearing credit that is not trading income, and the annual depreciation journal ($17,684 last July) has not yet been booked. The comparable underlying figure is closer to +$20,300 before depreciation. That is still a real improvement on July 2025 — and the improvement is cost discipline, not revenue growth. We would rather you read that here than find it yourself.

Every figure in this room is drawn from Artefact Group's own CRM and invoicing records as at 31 July 2026, or from a cited third-party source. Where a number is a projection or a recommendation rather than a record, it is labelled as such.

Why "events delivered" moved from 63 to 150. The old figure counted signed deals — one line for InStitchu's entire annual contract, whether it covered one expo or eighty. The corrected figure counts individually dated activations from our own delivery records: every InStitchu pop-up and expo-city stop (83 of them), every Brisbane Cycling Festival and ProVelo race stage run as its own operation, every discrete production across three years. Internal admin, website builds and one-off personal favours are excluded. Recounted 12 August 2026, from Fibery project delivery records — refreshed weekly from here.

What we actually sell

Four layers. Clients buy the top one and inherit the rest.

1 · Delivery

We run and build events end-to-end, white-label. Production, site and safety management, staging, hardware hire, warehousing, crew. This is the revenue engine and the reason we are on site at all.

2 · Instrumentation

Optic and Sense nodes, plus Vision reading whatever cameras the venue already owns. Counts, flow, occupancy, queues, heatmaps, dwell. Deployed as part of the build, not as a separate procurement.

3 · Interpretation

Aura and Pulse turn raw counts into decisions — where to move staff, when to open a second bar, whether the sponsor got the impressions they paid for. This is where the human logic gets encoded.

4 · Autonomy

The agent layer. Company-assigned MCP connections let back-office agents execute the admin — book the travel at the class the employee file permits, route it for approval, reconcile it overnight. The agents do the admin. The humans do the networking.

Why this is defensible

A dashboard is not a moat. Three things here are.

Negative-cost data acquisition

Competitors buy or beg for deployment sites. We are paid to be there. Every production contract is a sensor deployment, a labelled outcome, and a paying customer simultaneously. That is a structurally lower cost of data than any pure-play analytics vendor can reach.

Ground truth nobody else holds

We know the bump-in schedule, the crew roster, the bar layout, the weather call, the gate open time and what actually went wrong — because we ran it. A sensor vendor sees a count. We see the count and the decision that caused it. That pairing is the training signal.

Switching cost through operational entanglement

Once the agent layer holds a client's approval rules, supplier connections and site history, replacing it means rebuilding an operating system, not swapping a vendor.

And the honest counterpoint

None of that is proven at scale yet. As of today the data asset is thin, the interpretation layer is largely manual, and the agent layer is early. What follows is a plan to make the claim true, not a claim that it already is.

Why us

Operators first. That is the entire edge.

Every claim in this room reduces to one bet: that the people who physically run events are best placed to instrument them. We work with clients on a physically digital level — we build the site, run the show, and wire what the event already produces into one system. That only works if the founders are both things at once: operators and builders.

Wesley Pearson

Founder & Managing Director, Artefact Group · Founder, CrowdSense

Built the operating proof. Grew Artefact Group from $65k (FY23) to $527k (FY26) of won contract value across 150 individually delivered activations — production, site and safety, staging, crew — while designing the sensing platform in parallel.
Digital-first lineage. Built web analytics, ticketing platforms and data dashboards before instrumenting physical sites. CrowdSense is the same discipline pointed at physical space — the "physically digital" thesis is a biography before it is a strategy.
Hands on the hardware. Personally builds and field-tests the node fleet — edge-AI cameras, BLE counting nodes, LoRa mesh networking, battery telemetry. The specification in the CrowdSense tab is not outsourced.
Embedded in the regional event economy. Moreton Bay City Council is a top-five customer by recognised revenue; delivery partner on Innovate Moreton Bay alongside Creative HQ; in board-level conversations within the regional tourism and events ecosystem.

Jarrad Thessman

Managing Director, Midnight Assembly · Incoming co-founder, CrowdSense formalisation in progress

National events operator. Runs Midnight Assembly with delivery reach across the national festival and expo circuit, working alongside the Raft Projects crewing network out of Victoria.
Already in the trenches with us. Delivered alongside Artefact on One Fine Day Melbourne and the Melbourne Wedding Expo; has scoped CrowdSense for festival-scale deployment.
Product conviction, tested first. Joined the CrowdSense design-partner program as a user before the co-founder conversation started — conviction built on deployments, not a pitch deck.
Complementary networks. Artefact holds South East Queensland; Midnight Assembly's standing and relationships extend the reach nationally. Two operator networks, one data platform.

Why an M&A lens, while doing what we do best

Operators integrate operators

The graveyard of small-agency M&A is buyers who never ran an event. We run the same workflows daily — an acquired operator keeps its crew and its calendar, and gets instrumented from day one. Integration risk is lowest for the buyer who already does the work.

Deals come from the network

Small operators of recurring events sell to people they trust with their event, not to the highest spreadsheet. Two founders with live delivery networks — SEQ and national — see retirements and exits before they are ever listed.

The discipline holds

The criteria and preconditions in the Data via M&A tab still govern: buy longitudinal audience data, not revenue; nothing before the core is self-funding. Until then the founders' job is what it is today — deliver, instrument, compound.

The circle forming around the company

Status labels are literal: a conversation is listed as a conversation. Nothing here is claimed as revenue until it is contracted.

Pathway	What it is	Status
CrowdSense Pulse — Design Partner Program	Flat $3,995 ex per event: hardware, software, installation, on-site support and real-time insights. Limited partner slots, feedback-for-access.	live — partners onboarding
Untitled Group	LWM26 pedestrian counting delivered inside an existing relationship; design-partner interest in the wider program	delivered · in conversation
Tourism & Events Moreton Bay	CrowdSense across KiteFest and regional events; event-concept collaboration	in conversation
Creative HQ (NZ)	Delivery partner on Innovate Moreton Bay (active client); supporting a startup Lab build; launch support for CHQT — Creative HQ Queenstown	active client · expanding
Moreton Bay ↔ Queenstown JV	Proposed joint venture and social enterprise spanning the two innovation ecosystems, as a vehicle to build and develop CrowdSense	concept — in discussion
International showcase	Product review cycle with major-events advisors across Middle East and UK networks; building a global presence ahead of the 2032 decade	active conversations

The evidence

What four years of records actually show

Pulled directly from the Artefact Group CRM. Won contract value by close date, Australian financial years.

Won contract value FY2027 (begun 1 July 2026) is omitted — one closed deal at $20,000 is not a trend

Four charted years total $1,161,919 across 62 won deals. Add the single FY2027 deal and the all-time figures used elsewhere in this room — $1,181,919 across 63 won deals — reconcile exactly. That is a count of signed deals, not of activations delivered — see the Thesis tab for the corrected 150-activation delivery count.

Won contract value against recognised revenue

Two different measures of the same business. The gap is when the work is delivered, not a disagreement.

Won contract valueRecognised revenue (Xero)

Won value is dated by deal close; recognised revenue is Xero accrual accounting. FY2026 carries roughly $81k of won work into the following period.

Year	Won contract value	Recognised revenue	Gap
FY2025	$345,750	$301,369	$44,381
FY2026	$527,259	$446,002	$81,257

FY2026 won value

$527,259

31 deals · +52% on FY2025

FY23 → FY26 CAGR

~100%

$65k → $527k on won value

Win rate by count

51%

63 won / 60 lost, all time

Win rate by value

35.4%

we lose the large ones

Recognised revenue FY26 · Xero

$446,002

accrual, unaudited · FY25 $301,369 (+48%)

Net profit FY26 · Xero

$11,877

FY25 −$19,677 · first profitable year

Cash at bank

$66,963

as at 13 Aug 2026 · was $72,053

Working capital

−$13,491

AR $45.1k (58% overdue) · AP $28.8k (98% overdue)

Won contract value (CRM, by close date) and recognised revenue (Xero, accrual) measure different things — contracts are counted when won, revenue when earned. Both are shown; neither is audited.

Three years, one page

The financial record, and what FY2027 is budgeted to do

Two audited-basis years of Xero actuals, then a budget built line by line off them — not a target reverse-engineered from a raise. Every figure below reconciles to the accounting system on 5 August 2026.

Australian financial year	FY2025actual	FY2026actual	FY2027budget	FY2027stretch case
Trading
Incomerecognised revenue, accrual basis	$301,369	$446,002	$535,203	$624,403
Operating expenditure35 accounts, full chart	$321,045	$434,125	$469,027	$516,067

Net result	−$19,677	$11,877	$66,175	$108,336
Net margin	−6.5%	2.7%	12.4%	17.4%
Income growth on prior year	—	+48.0%	+20.0%	+40.0%
How the cost base behaves
Variable — moves with delivery volumeequipment hire, site and storage, contractors, event production, travel	—	—	$282,23660.2% of budgeted cost	$329,276
Fixed, scheduled or held flatdepreciation, insurance, utilities, telco, base wages	—	—	$186,79239.8% of budgeted cost	$186,792

Budget case assumes +20% income, award wages +4.75% from 1 July 2026, and public liability premiums +35%. Stretch scales only the variable portion of the cost base — fixed costs are held.

Why the second half is the number that matters

FY2026 did not improve gradually. The first half lost $36,150; the second half made $48,027 — an $84,177 swing inside one year, on a business that did not change what it sells. What changed was the mix of work and the discipline over what it costs to deliver. FY2027 is budgeted on the second half being the normal state, not the exception.

Why a mostly-variable cost base is a strength

Sixty per cent of the budgeted cost base moves with delivery volume, and the monthly record proves it — those accounts swing eight-fold or more between a quiet month and a build month. A quiet quarter costs far less than a busy one. That is why a lumpy revenue line has not produced a lumpy solvency problem, and it is why the downside case is survivable rather than fatal.

Source: Xero profit-and-loss, accrual basis, retrieved 5 August 2026, unaudited. FY2026 was reconstructed month by month and the twelve months sum exactly to the full-year totals. No CrowdSense product revenue is budgeted — there is no signed paying customer for it, and budgeting revenue we cannot evidence is the one thing that would make every other number on this page worth less.

FY2027, six weeks in — tracking against budget

Income to date

$46,667

44 of 365 days · pro-rated pace ~$64.5k

Expenditure to date

$18,919

pro-rated pace ~$56.5k — well under

Net result to date

$27,748

flattered — see note

Underlying net result

~$15,748

excludes a $12,001 ATO clearing credit

Six weeks of a 52-week events business is a thin sample — FY2026 itself lost money in H1 and made it all back in H2, so a straight-line pace comparison this early is a pace check, not a verdict. Income is running behind a naive pro-rated pace; cost discipline is running well ahead of it, consistent with FY2026's pattern of controlling spend before revenue lands. Depreciation (~$17,684 booked last year) has not yet been posted this year, so even the underlying figure above still overstates the true run rate. We would rather show a number that might narrow than wait until it looks better.

Xero profit-and-loss, accrual basis, 1 July – 13 August 2026, retrieved 13 August 2026, unaudited. This block refreshes weekly (see the automation note in the footer) — the FY2025/FY2026/budget/stretch columns above it are fixed-year figures and do not.

What is actually on paper

Every claim in this room that rests on a signed document, and what that document is. Verbal interest is excluded by design — it is tracked separately and it is not evidence.

Counterparty	Instrument	Executed	Status
AusCycling	Contractor Agreement — production and operation of Brisbane Cycling Festival and related events	15 May 2026	Signed & countersigned
Creative HQ / Innovate Moreton Bay	Master Services Agreement — marketing and content retainer	10 Feb 2026	Fully signed
Moreton Bay City Council	Purchase Order EDV001226 — Start Up World Cup AV and production, issued in two tranches	Jul 2026	PO issued
Check 1-2 Productions	Licence to Occupy — Geebung warehouse and storage	Aug 2025	Counter-signed, paying monthly
InStitchu	Recurring expo production, logistics and storage — invoiced and remitted across four expo cities	ongoing	Invoiced & paid

The honest boundary of this table

CrowdSense does not appear on it. The product has real commercial momentum — design-partner conversations in motion, an international review completed, a co-founder relationship being formalised — but as at 5 August 2026 it has no signed paying customer. Everything above is Artefact Group delivery work. That distinction is the whole reason for the raise: the services business is contracted and repeating, and the capital is to put the same ink under the product. A room that blurred those two would not survive an afternoon of diligence.

Commitment status is now a tracked field on every opportunity in the CRM — tier, signed-document link, signed date, and the single next action that moves it up a tier. Added 5 August 2026; the register behind this table is maintained there, not in anyone's memory.

Concentration is the diligence risk

Three accounts carry 72.3% of FY2026 won value. Across all four years, InStitchu alone is a third of everything ever won.

FY2026 recognised revenue, by customer

Five accounts carry 84% of the year. Xero recognised revenue, not won contract value.

InStitchuProVeloCheck 1-2AusCyclingMoreton Bay City CouncilEvery other client

Recognised revenue totals $446,002 for FY2026 — $81k below won contract value, which is timing, not a discrepancy.

Customer	FY26 revenue	Share
InStitchu	$151,011	33.9%
ProVelo	$101,565	22.8%
Check 1-2	$44,428	10.0%
AusCycling	$39,760	8.9%
Moreton Bay City Council	$38,415	8.6%
Every other client	$70,823	15.9%

Account	Won all time	Share	Cumulative
InStitchu	$398,257	33.7%	33.7%
AusCycling	$256,095	21.7%	55.4%
ProVelo	$196,617	16.6%	72.0%
HVIA / Brisbane Truck Show	$64,500	5.5%	77.5%
Innovate Moreton Bay	$55,595	4.7%	82.2%
M&C Saatchi S&E	$43,511	3.7%	85.8%
WSAA SuperExpo	$25,855	2.2%	88.0%
Greenfox Media	$25,000	2.1%	90.1%
Superdream	$22,550	1.9%	92.1%
Hop To It Collective	$21,144	1.8%	93.8%
Total won, all time	$1,181,919	100%

Collections, reconciled

Actual receivables per Xero as at 13 August 2026 are $45,060 across 11 invoices, 5 debtors — up from $25,238 two weeks earlier, entirely through new invoicing, not new delinquency: the $17,410 that is three-plus months old is the same aged balance as before, unchanged. 58.4% of the total ($26,327) is now overdue, down from 100%, because $18,733 of the increase is current, not-yet-due invoicing. The largest overdue line is still ProVelo at $12,531, unmoved since reconciliation began. InStitchu's exposure has grown from one current-cycle invoice to four — $7,828 now just past due, a further $15,771 across three invoices due 14 August — reflecting higher InStitchu volume, not a collection problem. Concentration and collection are related risks here — not, as the unreconciled register once suggested, the same client.

Reconciled against the Xero aged-receivables report, 13 August 2026 (first reconciled 31 July 2026). The CRM invoice register's non-paid lines materially overstate AR (an unreliable Account Type field plus stale lines) — Xero is the source of truth for receivables throughout this room.

CrowdSense: the finding that changes the plan

Sold on its own, CrowdSense does not close. Attached to delivery, it does.

Win rate by value — the same product, two ways of selling it

Share of pitched contract value actually won, four years of CRM records.

Attached rate is the whole-business win rate by value; standalone is the eleven analytics-only pitches. Unpriced losses are excluded from both.

Sales motion	Win rate by value	Detail
Attached to a delivery contract	35.4%	won inside production work
CrowdSense sold on its own	2.6%	$4,500 won · $169,500 lost

Standalone win rate

2.6%

by value · 1 win, 10 losses

Value lost

$169,500

standalone analytics deals

Lost standalone deal	Value
MBRIT — CrowdSense	$75,000
EarthCheck — CrowdSense	$35,000
QTIC — CrowdSense	$25,000
Ipswich City Council — crowd analytics	$10,000
RQS The Ekka — CrowdSense	$9,000
Wedding Expos Australia — lead scanners	$8,000
Sea Otter — CrowdSense	$7,500
Brisbane Powerhouse · Brisbane Festival · Moreton Bay City	unpriced

Where it does work

Brisbane Cycling Festival 2025 — roughly 27,500 visits and 22,500 unique attendees measured — was delivered inside the AusCycling production contract, and AusCycling has since become a $256k lifetime account across repeat years. Untitled Group's LWM26 pedestrian counters closed as an attachment to an existing relationship.

The strategic conclusion: CrowdSense is not yet a product business. It is the thing that makes the production business stickier, higher-margin and harder to displace — and the mechanism by which the data asset gets built. Selling it as standalone SaaS to councils has a four-year track record of failure. Stop.

The external evidence — the M&A market is paying for exactly this

While this room argues that events plus data compound, better-capitalised acquirers have been acting on the same thesis:

Transaction	Price	Stated driver
Cvent → ON24 (Dec 2025)	US$400m	Lead scoring, attribution
Cvent → Goldcast (Dec 2025)	~US$300m	Event content → marketing assets
Bending Spoons → Eventbrite	~US$500m	Audience and search
Truelink → GES (from Viad)	US$535m	Exhibition services consolidation
Encore → FIRST (Dec 2025)	undisclosed	"Data and insights" capability

Cvent alone deployed roughly US$700m in a single month. Informa — £4.04bn revenue FY2025 — attributes demand to "the value of proprietary First Party Data… and the deployment of AI technology." Sources, the acquisition criteria and the full thesis live in the Data via M&A tab.

Small services agency

3–5×

EBITDA multiple

Data / analytics

8–12×

the premium tier

What buyers pay the premium for is consistent across these deals: recurring revenue, proprietary first-party data, retention, and businesses that run without their founder. The entire Horizon 1–2 plan maps to those four. That is the design, not a coincidence.

The entrants validate the geography as well as the thesis: TBA Group launched in Australia in November 2025 explicitly citing Brisbane 2032, and Encore bought FIRST for its data capability. The window to build the local data position is open — and being noticed.

Why this market is a growth avenue, not just an exit backdrop

Multiple arbitrage. Small operators trade at 3–5× EBITDA. The same earnings, carried inside a data platform with recurring revenue and a transferable record, price toward the 8–12× tier. Buying at the low multiple and re-rating inside the platform is the single largest value lever available to a company this size.
Data has a time machine problem. A ten-year attendance history cannot be built at any price — only bought. Every year of waiting is a year of longitudinal data a better-capitalised entrant can buy first.

The supply side is a succession wave. Recurring consumer events are disproportionately held by owner-operators nearing exit with no succession plan. They sell off-market, to operators they trust with their event — not to the highest spreadsheet.
Integration is our home ground. An acquired event keeps its crew and calendar and gets instrumented from day one — the buyer already runs the same workflows. The classic agency-M&A failure mode (buyers who never ran an event) does not apply.

The deal register and relationship coverage can start now at near-zero cost — cash acquisitions still wait for the preconditions in the M&A tab. The full option, alongside five others, is laid out in Growth Strategy → Growth options.

The immediate problem

Against a $527k year, the live pipeline is two deals worth $8,495. There is no revenue-generating activity scheduled that would replace FY2026.

Meanwhile $241,500 sits dormant in Graveyard — including two deals that are neither lost nor dead:

Dormant	Value	Read
Oktoberfest (WEEZEN) 2026 — DISENYO	$150,000	reactivate
Brisbane Cycling Festival 2027 — AusCycling	$85,000	existing account
Paramount × CrowdSense (4 cities) — Superdream	$5,000	bundle it
Induction platform — safety services vendor	$1,500	low priority

Sequencing consequence

No growth strategy survives an empty pipeline. Everything in the Horizon 1 plan is subordinate to one thing: rebuilding forward cover to 2× the trailing year within 90 days. BCF27 at $85k is a warm renewal with a $256k account and should be the first call, not a strategic initiative.

Why the big ones get lost

51% of deals are won but only 35.4% of value. The losses cluster above $20k — CommBank SXSW $950k, The Ben Buckler $120k, AusCycling ops contract $80k, MBRIT $75k, Stadiums Queensland $50k, WaveCRM $50k, ALC Summit $45k. The pattern is consistent with competing on scope against larger integrated suppliers without a differentiated claim. That is precisely what the data asset is supposed to fix.

Market conditions

Good timing, tightening money

South East Queensland is entering the most concentrated decade of event investment in Australian history. It is also entering it with rising rates, flat marketing budgets and sharply rising cost-to-serve. Both are true and the strategy has to hold both.

The tailwind is real and local

2032 venue programme

$7.1bn

17 venues · construction began H1 2026

Qld visitor economy

$45.5bn

YE March 2026 · intl +35.2% YoY

Signal	Figure
Tourism & Events Qld grants, FY26-27	$96.25m +28% YoY
TEQ events visitor expenditure target	$1.080bn
Brisbane major events supported 2025	36 worth $315m
Future Brisbane events already secured	26 worth $295m
Brisbane events + performances delivered	928 / 2,036
BCEC events hosted, CY2025	853
Australian business events visitor spend	$17.2bn (CY2025)
ABEA forward intl pipeline	$743.5m delegate spend
…of which include an exhibition component	67%
Homegrown Destination Events Fund	$14m

Sources: Qld Govt venue funding · Delivering 2032 · Qld Budget SDS FY26-27 · BEDA 2025 review · BCEC · Tourism Research Australia · ABEA forward calendar · Homegrown fund

The headwind is equally real

RBA cash rate

4.35%

+75bp across 2026 · rising

Marketing budgets

+1.3%

2025→2026 · 7.8% of revenue

Pressure	Movement
Headline inflation, peak Jun-qtr 2026	4.8%
Business investment growth by Dec 2026	0.8% from 3.9%
GDP growth, Dec 2026	1.3%
Modern award wages from 1 Jul 2026	+4.75%
Public liability insurance vs pre-COVID	+30–50%
Venues / stages lost since COVID	1,300+
Australian ad market 2026 (WPP forecast)	+6.5% to A$30.7bn

What this means for pricing

Volume is not the constraint — Australian live events set revenue records in 2024 on record attendance. Margin is. Award wages up 4.75% and insurance up 30–50% land directly on event delivery cost, while client marketing budgets grow 1.3%. A pure labour-and-logistics business gets squeezed from both ends. A business that sells a measurable outcome does not.

Sources: RBA cash rate · RBA SMP May 2026 · Gartner CMO Spend 2026 · FWC Annual Wage Review 2026 · State of Australian Live Music 2026 · WPP Media

Competitive weather

International groups are arriving because of 2032

TBA Group — 40 years of Formula 1, FIFA World Cup and Olympic hospitality programmes — launched in Australia in November 2025 with a Sydney HQ and staff in Brisbane, explicitly citing the 2027 Men's Rugby World Cup, 2029 Women's Rugby World Cup and Brisbane 2032 as the reason. Encore acquired global brand-experience agency FIRST in December 2025, citing its “data and insights” capability.

Read: the thesis is validated by people with far more capital. The window to establish a defensible local data position is measured in a small number of years, not a decade.

Analytics vendors are consolidating, not commoditised

Xovis went PE-backed (Capvis majority, 2019) and became an acquirer itself, buying FORVIA HELLA's People Sensing business in 2024 — now 200+ staff and 120+ airports. The top five people-counting vendors hold only 35–40% of revenue, so the category is still fragmented. Hardware is 63% of that market's revenue and installation labour dominates five-year total cost.

Read: competing as a sensor vendor means competing on installed base and calibration labour against PE-funded consolidators. Competing as the operator who already owns the site access is a different, better game.

Numbers we are deliberately not quoting

Three published crowd-analytics market sizes for 2026 differ by up to 2.9× for the same year (US$4.84bn, US$1.99bn, US$1.64bn). No credible Australian crowd-analytics or experiential-agency market size exists in public sources. Rather than pick the flattering one, this room presents no analytics TAM. Any investor who wants one should be given a bottom-up build from SEQ venue counts and observable pricing — which we can produce on request.

Sources: TBA Group launch · Encore / FIRST · Xovis · Mordor Intelligence · Vendor pricing comparison

Entity architecture

One operating core, three commercial faces

The entities exist so that a services business, a data asset and a product can be capitalised differently without pretending to be the same thing. Investors price them differently — services on EBITDA, data and software on multiples of recurring revenue.

Artefact Group Pty Ltd trading

ABN 77 661 296 715 · the engine

End-to-end white-label delivery: production, site and safety, staging, hardware hire, warehousing, crew, content. Holds the client relationships, the balance sheet and the operating history. Generates the cash and — critically — the deployment access that everything else depends on.

CrowdSense product

Aura OS, Pulse, Vision, Optic and Sense nodes. Today an attached capability inside delivery contracts. The plan is to earn its way to a separate entity once it carries recurring revenue on its own — not before, because the evidence says a premature separation is what produced a 2.6% standalone win rate.

AdmitMe dormant

Ticketing / registration / access

Ticketing, registration and exhibitor portals. Currently inactive in the pipeline. Its strategic value is as the identity layer — the join key between a ticket, a person and a movement trace. That is what makes the data asset saleable rather than anonymous.

The structural argument

Delivery gives us the site. Ticketing gives us identity. Sensing gives us behaviour. Interpretation gives us meaning. Any one alone is a commodity; together they are a closed loop that no pure-play vendor in this market currently holds. The corporate structure should follow that logic — hold the IP where it can be capitalised, and licence it into the trading entity on arm's-length terms so the data asset's value is legible to an acquirer.

Recommended sequencing

Step	When	Why
Keep everything inside Artefact Group	Now → H1	Below the threshold where separation adds anything but cost and complexity. Fix the pipeline first.
Formalise the IP register and licence terms	H1	Cheap, reversible, and materially improves how a diligence process reads the data asset. Do it before you need it.
Incorporate CrowdSense separately	On ~$250k ARR	Only once recurring revenue exists to capitalise. Premature incorporation creates a shell with a story and no numbers.
Reactivate or divest AdmitMe	H2	Either it becomes the identity layer or it is a distraction. Decide explicitly rather than letting it drift.

This section is a recommendation, not a record. Confirm structure and any IP-licensing arrangement with your accountant and lawyer before acting — there are tax and Division 7A considerations that sit outside what this document can assess.

Growth strategy

Three horizons, one dependency chain

Each horizon funds the next. Horizon 1 is not strategy — it is survival arithmetic, and nothing else starts until forward cover is rebuilt.

Horizon 1 · now

Refill and stabilise

Aug 2026 → Jan 2027 · target: $1.0m forward cover

Reactivate the graveyard. BCF27 ($85k, existing $256k account) and Oktoberfest WEEZEN ($150k) are calls, not campaigns. $235,000 between them — just under a quarter of the $1.0m forward-cover target, from two phone calls.
Collect and enforce. Reconciled AR is $25k — every dollar of it overdue, $17k by three-plus months. Chase ProVelo and Check 1-2, make deposit-and-progress billing standard on every new contract, and keep the register reconciled monthly so the books never drift 5× from reality again.
Attach analytics to every production contract as standard inclusion, priced into the job rather than quoted separately. Stop pitching it as a line item that can be declined.
Renew the anchors early. AusCycling, InStitchu and ProVelo are 72% of FY26. Multi-year terms with indexation, signed before the counterparty goes to market.
Hire one salesperson. Not a marketer. The pipeline problem is a coverage problem and there is currently no one whose job it is.

Horizon 2 · next

Productise the loop

Feb 2027 → mid 2028 · target: $250k recurring

Turn attached analytics into a subscription. Venues and recurring festivals move from per-event fees to an annual seat + site licence. This is the first genuinely recurring line.
Ship the benchmark. Once ~20 comparable sites are instrumented, the product stops being “your numbers” and becomes “your numbers against everyone like you.” That is the first thing councils and tourism bodies could not refuse — and the reason the earlier pitches failed.
Package the agent layer as a paid operations tier for clients who already trust us with delivery. Travel, approvals, supplier reconciliation, induction and compliance.
Build the warehouse into a hire P&L. Infrastructure hire is capital-efficient revenue that fills the gaps between production peaks and improves utilisation of assets already owned.
Second and third anchor sectors beyond cycling and wedding expos — motorsport, agricultural shows, and the 2032 venue test-event programme.

Horizon 3 · later

Compound and consolidate

2028 → 2032 · the Games decade

Acquire for data, not revenue. Small operators of recurring consumer events hold irreplaceable longitudinal audience data. See the M&A tab for the specific thesis.
Become the measurement standard for SEQ public events ahead of 2032, when every venue and council will be required to evidence attendance, dwell and economic contribution.
Licence the model outward — sell interpretation to operators we do not deliver for, once the benchmark is thick enough to be valuable without our crew on site.
Position for the structural exit. Event tech M&A is explicitly data-driven and data/analytics agencies trade at 8–12× EBITDA against 3–5× for small traditional shops. The whole strategy is about which of those two multiples applies.

The one-page logic

Horizon	Primary metric	Target	Fails if
H1 — Refill	Forward-booked contract value	$1.0m by Jan 2027	No dedicated sales capacity is hired
H1 — Collect	Aged receivables > 60 days	< $25k	Anchor client relationship is prioritised over payment terms
H2 — Recur	Annual recurring revenue	$250k by mid-2028	Analytics stays a per-event add-on
H2 — Instrument	Comparable sites in benchmark	20+	Deployments aren't standardised enough to compare
H3 — Compound	Share of revenue that is data or software	> 30%	Acquisitions are made for revenue instead of data
H3 — Concentrate	Largest client share of revenue	< 20%	Growth comes from existing anchors only

Targets are recommendations set against the observed base, not forecasts derived from a financial model. Treat them as the thresholds at which each horizon is judged to have worked.

Growth options — six ways to compound from here

The three horizons are the spine. Around them sit six distinct growth engines — some running, some optional. Each opens into a full plan: why it works from this exact starting position, the steps, the capital, the effect on the multiple, and the criteria that would kill it.

Option	Engine	Capital intensity	First revenue	Multiple impact	Status
1 · Organic core	Deliver + instrument, renew	Low — funded by pre-seed	Now	Gradual re-rate	in motion
2 · Data-led M&A	Buy audience history, instrument it	Low cash if earn-out / vendor-financed	2027+ · register starts now	Step-change	criteria set
3 · Design partners → benchmark SaaS	Product recurring revenue	Medium — engineering + site ops	Q1 FY27	The re-rate itself	cohort onboarding
4 · ANZ corridor JV	Two ecosystems + impact funding	Low — ring-fenced establishment	2027	Optionality	concept — in discussion
5 · International lighthouse	District-scale deployments	Medium — travel + localisation	One deal away	Reference premium	active conversations
6 · White-label network	Partner distribution	Very low — documentation + support	H1 FY27	Recurring via channel	structure being shaped

These are options, not six simultaneous priorities. Options 1 and 3 are the plan of record; the others are sequenced behind explicit gates — each modal states its own.

CrowdSense · powered by Artefact Group

Know the crowd. Craft the moment.

Five modules across hardware, vision and analytics. Counts and events leave the device; footage does not. That single architectural decision is what makes the data asset legally durable.

Aura OS software

Live operational dashboard

Real-time counts, occupancy gauges and threshold alerts during the event, when a decision can still change the outcome.

Pulse software

Post-event analytics

Forecasts, recommendations and funder-ready reporting. The artefact a council or sponsor needs to justify the spend.

Vision software

Camera-based analytics

Runs against CCTV the venue already owns — stated support for 3,600+ camera brands. Zero-capex entry into an existing site.

Optic Nodes hardware

Edge-AI counting cameras

Precise counts at gates and chokepoints. On-device inference, so the count leaves and the footage stays.

Sense Nodes hardware

Crowd pressure sensors

Perimeter density sensing — the safety-relevant measure that gate counts alone cannot give you.

What it measures

Across all five modules

People counting · visitor flow · occupancy · queue detection · heatmaps · vehicle tracking · anonymised audience profiling · anti-tailgating.

Optic Node — specification

Camera	4 MP · 2688 × 1520
On-device AI	0.6 TOPS · 25 fps detection
Lenses	51° · 88° · 137° field of view
Connectivity	Wi-Fi 6 · 4G LTE · Ethernet with PoE
Power	AA batteries · USB-C · PoE
Environmental	IP67 · −20 °C to 50 °C
Dimensions	77 × 77 × 48 mm
Local storage	microSD to 2 TB
Certification	CE · FCC
Transport	Encrypted MQTT / HTTPS
Integration	REST API · webhooks
Firmware	OTA updates with rollback
Duty cycle	Motion-triggered — months to years on battery

Source: CrowdSense 2026 deck, retrieved 31 July 2026.

Privacy is the architecture, not a policy

The design decision that matters

“Footage stays on the device — counts and events are what leave.” Inference happens at the edge. What crosses the network is a number, not a person.

This is not a compliance footnote. It is the difference between a data asset that survives regulatory scrutiny and one that becomes a liability the moment a regulator or an acquirer's counsel looks at it.

Counts and events are inherently anonymous — there is no biometric template to leak, and no footage in transit to intercept.
Audience profiling is described as anonymised, which materially changes the Privacy Act position versus identified profiling.
Ariadne's positioning of camera-free sensing as “structurally outside EU AI Act Annex III high-risk biometric categories” shows where regulation is heading. Edge inference with no footage egress is a defensible position on the same axis.

The caveat to hold onto

The moment movement data is joined to ticketing identity — the AdmitMe thesis in the architecture tab — this protection weakens considerably. That join is where the commercial value concentrates and where the risk concentrates. Take formal privacy advice before commercialising joined datasets, not after.

Proven, prototyped, roadmap

An investor's technical advisor will ask exactly this. Answering it precisely is worth more than implying everything ships today.

Capability	Status	Evidence
Visitor analytics at a major public festival	Delivered	Brisbane Cycling Festival 2025 — ~27,500 visits, ~22,500 unique attendees. Client testimonial from Mathew Hey, AusCycling, citing real-time crowd density and movement pattern monitoring
Pedestrian counting as a contracted deliverable	Delivered	Untitled Group, LWM26 — closed and delivered as an attachment to an existing relationship
Edge-AI counting nodes	Productised	Specification published; deployed in field conditions. Deployment count should be disclosed precisely in diligence
Camera-brand-agnostic vision layer	Delivering · expanding	In delivery against venues' incumbent camera stock (3,600+ brand vendor stack); integration coverage expanding site by site — the exercised-integration register records exactly which have run live
LoRa mesh for off-grid multi-node sites	Delivering · expanding	In delivery and expanding: Meshtastic mesh with ESP32-based PAX counting nodes; RAK3312 / WisMesh field enclosures (MHF4-to-IPEX antenna bridging); INA219 battery telemetry in firmware
Camera zone-handoff logic	Delivering · expanding	In delivery and tuned in the field: CamThink NeoEyes NE301 deployed alongside Milesight X5 — the strongest field sensor for off-grid MQTT deployments
MQTT + Grafana telemetry pipeline	Delivering · growing	Live telemetry pipeline running in delivery, growing with each instrumented site
Cross-site benchmark dataset	Roadmap	The Horizon 2 gate. Requires ~20 comparable instrumented sites before it has commercial meaning
Recurring site licences	Roadmap	No recurring analytics revenue exists today. This is the single most important thing to prove

What we would want asked

How many nodes are deployed right now? Across how many distinct sites? How many cumulative event-days of data exist? What is the measured counting accuracy against a manual control, and at what crowd density does it degrade? These are the questions that separate a real sensing business from a specification sheet — and the growth in those numbers is the most credible progress metric this company has.

Deployment modes

Temporary · event

Days · battery or generator

Nodes go in with the build and out with the bump-out. The default mode today, and the one that generates data as a by-product of production revenue.

Semi-permanent · season

Weeks to months · PoE or mains

Recurring venues and seasonal precincts. The bridge to recurring revenue and the first place a site licence makes sense.

Permanent · always-on

Continuous · fixed infrastructure

Venues, precincts and councils. Where longitudinal data actually compounds and the benchmark gets its depth.

Off-grid temporary deployment is the hardest of the three and the one competitors serving airports and retail are least optimised for. It is also where Artefact's production presence is an outright structural advantage: the crew, the power and the site access are already there and already paid for.

For producers & venue ops

See your crowd before it's a problem

CrowdSense shows you how many people are on your site right now, zone by zone, gate by gate. You get it live while the event's running, and a plain report once it's over.

Written for the people who run the ground

No investor language in this tab — it's the same walkthrough we'd give you on a site visit. If a term trips you up, the plain-English glossary at the bottom has you covered, and Plan my event below gives you an indicative kit for your site in under a minute.

If you're a producer

Your site exists for a few days a year, and it's rarely near mains power or reliable signal. Our BLE and camera nodes run on battery with solar top-up, and mesh together over LoRa, so you get one picture of the site even when the phone network can't reach every gate. Bump-in is quick — mount the nodes, power them on, and they find the mesh themselves — so your crew isn't buried in extra cabling. From the control room you'll see live counts at every gate and zone, so you know where the crowd is building before it's a problem.

If you run a venue

You've probably already got CCTV covering your gates and floor — Vision reads footage from it directly, across a wide range of existing camera brands, so you're not buying a second camera system to sit on top of the first. Because you run events week in, week out, Aura gives your control room the same live occupancy view for every show, not just the big ones. Pulse turns each event into a report you can hand to your safety officer or council afterwards — the record you want on file for duty of care. And where your cameras don't reach, our own nodes drop into the same system.

What event day looks like

Live in the control room

Aura shows you occupancy across your whole site as it happens — total numbers, and the breakdown by gate and zone. It's built to be read at a glance by whoever's on shift, not just the person who set it up.

Zones and alerts

Set a capacity for each zone, and Aura flags it as density climbs towards that line. That gives your crowd safety team time to open another gate, ease the flow, or hold entry — before it turns into a squeeze.

After the event, with Pulse

Once the gates close, Pulse turns the day's data into a report — ingress and egress over time, peak occupancy by zone, and how your crowd actually moved. It's the record you keep for your own debrief, and the one you hand over if a safety officer or council asks.

What your control room sees

An illustrative run of one show day in the Aura view — how a site fills, where density builds, and when the alerts fire. This is a simulation to show you the shape of it, not live data from a site.

10:00show day

18on site now

18through the gates

18peak so far

Gate A3/min

comfortable up to 19/min

Gate B3/min

comfortable up to 13/min

Main arena8

0% of 8,000 capacity

Food & bar6

0% of 4,500 capacity

Second stage4

0% of 5,200 capacity

Alerts as they fire

Quiet so far — the day hasn't started.

How a deployment runs

Site walk & plan4–6 weeks out

Wesley or someone from the team walks your site with you — gates, pinch points, power, and where signal drops out. From that we plan camera and node placement in our site planner, and work out how the mesh will sit together.

Kit & config2–3 weeks out

We configure the cameras and nodes for your zones and set up Aura with your site map and capacity thresholds. Anything that needs council or safety sign-off gets sorted before kit arrives on site.

Bump-in and mesh-up1–2 days out

Nodes go up as part of your normal bump-in — on trussing, poles, or existing fixings. We power them on, confirm the mesh is talking end to end, and check every zone is reporting before doors open.

Showday live opsevent day

Your control room watches occupancy and density live through Aura for the whole event. We're on call throughout, and usually on site or nearby for a first event with us.

Debrief & Pulse report1–2 weeks after

We pack down the nodes with you, then Pulse turns the day's data into a report — occupancy, flow, and any alerts through the event. We sit down with you afterwards to talk through what worked and what to change next time.

One festival day, hour by hour

How a typical showday runs with CrowdSense on site — no invented numbers, just the rhythm of the day.

06:00Crew arrives and the node fleet comes off the truck. Batteries go on charge while trussing goes up around the site.

07:30Cameras and BLE nodes go up on gates and inside key zones, and we watch the mesh come together, node by node.

09:00Full system check before anyone’s near the gates — every zone reporting into Aura, alerts set to today’s capacity thresholds.

11:00Gates open. Control room watches ingress climb gate by gate, so you know early which entries are moving people through and which aren’t.

13:00Peak ingress. Busiest the gates get all day — as much attention on queue-side density as on the count itself.

17:00Afternoon lull. A good window to check node battery and confirm the mesh is solid before the evening crowd builds again.

20:30Headline set. Site occupancy is at its highest for the day, and density is the main thing on screen, especially front of stage.

23:00Egress. Exit gates and Aura both get watched closely as the crowd leaves — often the part everyone cares about most.

00:30Midnight wrap. Site’s clear, nodes come down or go to standby, and today’s data is already sitting in Pulse ready for the debrief.

Plan my event

Four quick questions, an indicative kit. Every site is different — a site walk settles the real plan.

What kind of event are you planning?

Roughly how many people are you expecting?

What’s the power situation on site?

Does the site already have CCTV in place?

Indicative only — every site, layout and event brief is different, and a short site walk is what turns this into a real plan.

"But we already…" — fair pushbacks

We already count with clickers or security head counts.

That's a reasonable start, and we're not knocking it. The difference is a clicker gives you one number at the gate. CrowdSense gives you live counts by zone, so you can see density building in a specific spot before it's a squeeze — and a report afterwards you can actually hand someone.

Our site has no power.

That's normal for a lot of the sites we work on. Our nodes run on battery with solar top-up and mesh together over LoRa, so they're built for off-grid sites, not bolted on assuming you've got mains everywhere.

We can't afford new cameras.

You might not need any. If you've already got CCTV, Vision reads footage straight from it, so the cost is the software layer, not a new camera system. Where there are gaps, we can talk about whether dedicated nodes make sense for just those spots.

Our patrons will worry about privacy.

A fair thing to ask, and worth explaining to them too. Our cameras count people and movement on the device itself, with no facial recognition and no identifying anyone — just counting how many people are where.

We only run two events a year.

That's fine — plenty of the producers we talk to are seasonal, not year-round. It's worth a conversation about what makes sense for your event size and frequency, rather than assuming you need a big permanent setup.

Your first event with us

Have a first call with Wesley about your event and site.
Send us your site map, capacity, and gate layout.
Book a site walk to plan camera and node placement.
Agree zones, capacity thresholds, and who watches the control room.
We configure Aura with your site map and thresholds.
Kit arrives and gets tested before it goes near site.
Nodes go up and mesh together during bump-in.
Confirm every zone is live before gates open.
Run the event with live occupancy in your control room.
Debrief with us and walk through your Pulse report.

Questions producers actually ask

What if our site has no power?

That’s normal for a lot of the sites we work with. Our nodes run on battery with solar top-up, and the LoRa mesh means they don’t need mains power or a phone signal to talk to each other. We’ll still ask about your event length, because that shapes how we plan battery and charging.

What happens if it rains?

The cameras and nodes are IP67 rated, so they’re built to handle rain, dust, and sitting outdoors for the run of your event. We still mount them sensibly — out of standing water and away from anything that could fall on them.

Are you filming and identifying our patrons?

No. The cameras count people and movement on the device itself, so the image never has to leave the camera for us to get a number. We’re not doing facial recognition and we’re not identifying individuals — we’re counting how many people are in a zone and which way they’re moving.

We already have CCTV — do we need new cameras?

Often not. Vision is our software layer that reads footage from your existing cameras, across a wide range of brands, and turns it into occupancy and flow data without you buying a new camera system. If there are gaps in your coverage, we can talk about filling them with our own cameras or BLE nodes.

What if the phone network drops out on the day?

That’s exactly what the LoRa mesh is for. Nodes talk to each other and back to the control point over their own mesh network, so you’re not relying on a telco signal that a big crowd can easily overload.

How long does bump-in take?

It depends on your site — the number of zones, how spread out your gates are, and what’s already there to mount to. As a rough feel, it slots into the rest of your bump-in schedule rather than needing a day of its own, but we’ll give you a proper estimate once we’ve walked the site.

What does the control room actually see?

Aura gives you a live view of occupancy by zone and gate, with alerts as a zone’s density climbs towards the capacity you’ve set. It’s built to be read at a glance by whoever’s on shift, not just the person who configured it.

Can our safety officer or council get access?

Yes, and it’s easy to set up. We can give a safety officer or council contact their own live view into Aura, or send them the Pulse report after the event — whichever your event needs.

What happens if a node drops out mid-event?

We monitor node health and battery through telemetry, so we usually know before you do. Because nodes mesh together, losing one doesn’t take down the whole picture — you lose visibility of that one spot until it’s back, not the whole site.

What do we get after the event?

A Pulse report — ingress and egress over time, peak occupancy by zone, and the flow patterns across your site. It’s built to be useful for your own debrief, and to hand to a safety officer or council if they ask for evidence of how the crowd moved.

Is our event too small for this?

Probably not — we work with sites from small community events up to major festivals, and it’s worth a conversation either way. The honest answer is it depends on your site and what duty-of-care evidence you need, so get in touch and we’ll tell you straight if it’s not a fit.

How do we get started?

The easiest way in is a pilot — one event, so you can see how it works on your own site before committing to more. Get in touch, we’ll walk your site or talk through your venue, and go from there.

Plain-English glossary

Edge-AIThe camera works out the count itself, right there on the device, instead of sending video off to a server to think about it.

PAX countingCounting people — PAX is just an events industry word for everyone moving through your site, gate, or zone.

BLE nodeA small battery-powered device that uses Bluetooth signals to estimate how many people are nearby, handy for spots a camera can’t cover.

LoRaA radio technology that sends small bits of data a long way on very little power — it’s what lets our nodes talk without wifi or phone signal.

Mesh network (Meshtastic)Instead of each node needing its own link back to base, they pass data to each other, so the network keeps working even if one link drops out.

MQTTA lightweight way for our devices to send their data back to one place reliably, without needing a constant heavy connection.

GrafanaThe dashboard we use behind the scenes to watch how the system itself is running — battery levels, signal strength, node health.

Zone handoffThe logic that makes sure someone walking from one camera’s patch into another gets counted once, not twice.

Ingress / egressPeople coming in and people going out — your gate counts, in both directions.

DensityHow packed a space feels, not just how many people are in it — the same headcount feels very different in a small pen versus an open field.

OccupancyHow many people are on your site, or in a particular zone, right now.

Vision layerOur software that plugs into cameras you already own and turns their footage into occupancy and flow numbers, without new hardware.

TelemetryThe health data your nodes send back about themselves — battery level, signal strength — so we know if something needs attention.

Duty of careYour responsibility to keep people at your event safe. The numbers and reports we give you help you show you took it seriously.

Talk to us

If you've got an event coming up, have a chat with Wesley about what CrowdSense would look like on your site.

Email Wesley about your event hello@artefactgroup.au

Or just reply to whoever sent you this room — it lands with the same two people either way.

— opens straight here, no wizard.

Go-to-market

Sell the outcome, include the instrument

The four-year record is unambiguous: analytics quoted as a separate line gets declined; analytics included in a delivery contract gets used, and the account renews. Every motion below is built on that.

Retire this motion

Cold outbound to councils and tourism bodies offering crowd analytics as a standalone purchase. Ten attempts, one small win, $169,500 of lost value, four years. It is not a messaging problem — a council has no budget line for it, no internal owner, and a procurement process that favours incumbents. Re-enter this segment only in Horizon 2, and only with a benchmark they cannot get elsewhere.

Four motions, ranked by evidence

Motion	Target	Wedge	Evidence	Priority
Land as producer, expand as platform	Sporting bodies, festival owners, expo operators	Win the production contract; instrument it as standard; return with a data-led renewal that is hard to compete against	AusCycling: $3.6k first deal (2023) → $256k lifetime across BMX, TrackNats and BCF, with analytics delivered inside the contract	Primary
Agency white-label	Creative and brand agencies without production capability	Be the invisible delivery arm. They keep the client; we take the build, the crew and the measurement	Superdream: 6 engagements FY26 (Scream7, guestHaus, Scary Movie 6, Paw Patrol live). M&C Saatchi $43.5k across CBA work	Primary
Multi-city rollout contracts	Brands running repeating national programmes	One contract, many cities, standardised kit and measurement. Highest revenue per unit of sales effort in the whole book	InStitchu: $398k lifetime across 83 expo activations FY24–FY26 alone, four consecutive annual renewals	Primary
Infrastructure & warehouse hire	Other producers, venues, councils	Sweat assets already owned between production peaks. Low sales cost, immediate margin, no new capability required	Hardware, warehouse and storage already on the balance sheet; InStitchu storage contracts already billed as a separate line	Quick win

The pricing change that matters

Today most work is quoted per event, at a price that has to absorb a 4.75% award increase and 30–50% higher insurance while the client's marketing budget grows 1.3%. That is a losing trade repeated annually.

Move to three tiers

Tier	What it is	Commercial shape
Deliver	Production, site & safety, crew, build	Project fee, indexed annually to the award increase — written into the contract, not renegotiated
Deliver + Measure	Above, with instrumentation and a post-event intelligence report as standard inclusion	Project fee at a higher rate. Never itemised separately, so it cannot be cut from the scope
Always-on	Permanent instrumentation, benchmark access, agent-layer operations	Annual site licence + seats. The recurring line, and the one that changes the valuation multiple

Indexation clauses are the single cheapest margin protection available and cost nothing to insert at renewal. Given award wages rose 4.75% on 1 July 2026, every contract renewed without one is a real-terms price cut.

First 90 days, concretely

#	Action	Value at stake
1	Call AusCycling on BCF27 — warm renewal, existing $256k account	$85,000
2	Reopen DISENYO on Oktoberfest WEEZEN 2026	$150,000
3	Chase overdue AR — ProVelo $12,531 · Check 1-2 $5,882 · InStitchu $7,828 reconciled 31 Jul · refreshed 13 Aug	$26,327
4	Convert Paramount × CrowdSense from graveyard by bundling into the Paw Patrol production already in play with Superdream	$10,500
5	Put multi-year, indexed terms in front of InStitchu, AusCycling and ProVelo before they go to market	~$380,000/yr
6	Write the standard analytics inclusion into the production quote template	structural
7	Hire one business development lead with events-sector relationships	the constraint
	Identified, addressable, already-warm	$271,827

The point

Nothing in the first 90 days requires new product, new market or new capital. It requires someone to make the calls. That is the honest read of where this business is: the growth problem is a capacity-to-sell problem before it is anything else.

The agent layer

The agents do the admin. The humans do the networking.

Events run on an enormous volume of low-judgement coordination — bookings, approvals, inductions, reconciliations, call sheets. It is the least valuable work in the business and it consumes the people whose relationships actually generate revenue. That is the arbitrage.

How it works in practice

Each supplier gets a company-assigned MCP connection. The employee file holds the entitlements — ticket class, room class, hotel star rating, spend ceiling. An agent takes a plain request, resolves it against those entitlements, executes the booking, routes it for approval and reconciles overnight.

The person asks for a flight. They do not open a booking tool, look up a policy, or file an expense. The policy is the context the agent already holds.

Why this is a product, not an internal efficiency

We already integrate first-party hardware, software and custom web apps into a client's operation. Teaching the system how a specific client operates — their approval chains, their suppliers, their site rules — is work we are doing anyway during delivery. Packaging it as an operations tier turns a cost of sale into a recurring line, and turns a delivery contract into an operating dependency.

Where the leverage actually is

Function	Today	With agents
Crew travel & accommodation	Manual, per person, per event	Requested in plain language, booked to policy, approved by exception
Contractor onboarding & inductions	Documents chased by email	Issued, tracked and escalated automatically; compliance is the default state
Supplier reconciliation	Post-event, from memory and paper	Matched against the project as costs land
Post-event reporting	Weeks later, hand-assembled	Assembled from sensor + roster + finance data on the night
Quote generation	From scratch each time	Priced from what comparable sites actually cost us

The connection back to the thesis

Every one of these agent actions is also a labelled data point about how an event is really run. The operations layer is not a side business — it is the mechanism that captures the human logic the interpretation layer needs. Automating the admin is how the collective brain learns.

Maturity, stated honestly

Parts of this are live and parts are design intent. For diligence, be precise about which integrations are in production today, which are prototyped, and which are roadmap — an investor will ask, and the answer is more credible than an implication that all of it ships now.

Data acquisition through M&A

Buy the audience history, not the revenue

The market is already paying for exactly this. The question is whether we are a buyer of small longitudinal datasets now, or a seller of one later.

The market has made the argument for us

Transaction	Price	Stated driver
Cvent → ON24 (Dec 2025)	US$400m	Lead scoring, attribution, demand gen
Cvent → Goldcast (Dec 2025)	~US$300m	Turning event content into marketing assets
Bending Spoons → Eventbrite	~US$500m	Take-private; audience and search
Truelink → GES (from Viad)	US$535m	Exhibition services consolidation
Encore → FIRST (Dec 2025)	undisclosed	“Data and insights” capability, 15,000 events/yr
Maritz → Convention Data Services	undisclosed	Registration and lead-capture data

Informa — the largest listed comparable, £4.04bn revenue FY2025 — attributes demand directly to “the value of proprietary First Party Data… and the deployment of AI technology.” Cvent alone deployed roughly US$700m in a single month.

What that implies for a business this size

Small agency

3–5×

EBITDA, <$1m EBITDA

Data / analytics

8–12×

premium tier

Acquirers pay the premium for recurring revenue above 80%, EBITDA margins above 20%, client retention above 90% and low owner dependency. Artefact today has none of those four. Every item in the Horizon 1 and 2 plan maps to one of them — that is not a coincidence, it is the design.

The strategic question in one line

Two businesses with identical revenue, one trading at 4× and one at 10×. The difference is recurrence, retention and whether the asset walks out the door with the founder. That gap is worth more than any single year's growth.

Acquisition criteria

Targets are small operators of recurring, consumer-facing, tech-enabled activations — the environments where the same audience returns annually and the operator has never done anything with the record of it.

Buy if it has

Three or more years of continuous attendance or ticketing history for the same event
A recurring calendar position — an annual show, not a one-off
An owner-operator nearing exit with no succession
Data we cannot generate ourselves at any price, because the history already happened
A venue or council relationship that is difficult to originate cold

Walk away if

The value is the revenue rather than the record — we can win revenue with a salesperson, more cheaply
Attendance data is anonymous with no join key to identity
Consent and privacy provenance is unclear — under the Privacy Act this contaminates the asset rather than adding to it
The event depends on the departing owner's personal relationships
It requires debt at a 4.35% and rising cash rate for a non-cash-generative reason

Sequencing discipline

This is a Horizon 3 activity. Acquiring anything while the live pipeline is $8,495 and working capital is negative would be a governance failure, not a strategy. The precondition is a stable, self-funding core — earliest realistic window is 2028.

Sources: Event Tech Live on Cvent · Skift on Eventbrite · Informa FY2025 results · FE International agency multiples

Value over time

What the multiple is actually worth

The whole strategy reduces to one question: does revenue growth arrive with recurrence attached, or without it? Move the sliders — the gap between the two curves is the entire argument for doing the hard version.

Starting revenue — FY2026 won contract value

$527k

Annual revenue growth

35%

Recurring share of revenue by 2032 — the strategy variable

35%

EBITDA margin

18%

Revenue 2032

$3.19m

at the growth rate set above

Blended multiple

6.1×

EBITDA, weighted by recurrence

Enterprise value — with recurrence

$3.5m

the strategy executed

Services-only comparison

$2.3m

same revenue, 4× multiple

Building recurrence to 35% is worth $1.21m more than the same revenue delivered as pure services — 1.5× the enterprise value on identical turnover. That difference, not the growth rate, is what the strategy is for.

Enterprise value — recurrence built Enterprise value — services only

Read this as a sensitivity, not a forecast

This is a multiple-driven illustration, not a financial model. It applies published agency EBITDA multiples (3–5× small services, 8–12× data/analytics) to a revenue path you set by hand. It does not model cash, working capital, headcount, capex or receivables. Its only purpose is to show the shape of the trade-off between growing revenue and changing the kind of revenue. Any figure taken into a term sheet needs a real financial model behind it.

Value milestones — what has to be true, and when

By	Milestone	Evidence it worked	Unlocks
Jan 2027	Forward cover rebuilt	$1.0m contracted, >3 months ahead	Ability to plan and hire at all
Jun 2027	Anchors on multi-year indexed terms	InStitchu, AusCycling, ProVelo signed beyond 12 months	Revenue predictability; removes the single biggest diligence objection
Dec 2027	First recurring analytics licences	>$100k ARR from site licences	The multiple conversation starts
Jun 2028	Benchmark live across 20+ comparable sites	Product sells without our crew on site	Re-entry to the council and tourism-body segment that has rejected us ten times
2029	Largest client below 20% of revenue	Concentration risk retired	Institutional capital becomes available
2030	First data-led acquisition	Longitudinal dataset acquired and integrated	Compounding beyond organic delivery capacity
2032	Measurement standard for SEQ public events	Games-adjacent venues and councils using the benchmark	Strategic rather than financial exit

Implied value — three lenses, and which one is the wrong instrument

Investors ask what the business is worth today. The honest answer is that the two lenses which are easiest to calculate are both the wrong ones, and saying so is more useful than picking the flattering number.

1 · Trailing earnings multiple

FY26 EBITDA proxy is roughly $29.6k (net profit plus depreciation, no interest expense). At the published 3–5× small-services band that implies $89k–$148k.

Why it is the wrong instrument: it values a business that has just spent four years buying capability — $104,616 of deployed hardware sits on the balance sheet — as though this year's thin margin were its steady state. Trailing multiples price maturity. This is not that.

2 · Asset and replacement floor

Net equity is $42,408 against total assets of $235,373, including $104,616 of owned production and sensing hardware, plus four years of client relationships and a working CrowdSense stack.

Why it is a floor, not a value: it says what it would cost to rebuild the kit, not what the kit earns. Useful as downside protection in a term-sheet conversation; useless as a growth valuation.

3 · Forward, recurrence-weighted

The interactive model above is this lens. It is driven by one variable that trailing multiples cannot see: the share of revenue that recurs, which moves the applicable band from 3–5× toward 8–12×.

Why it is the relevant one: a pre-seed is priced on what the next three years can be made to look like, and the whole strategy is a plan to move one number. Set the sliders yourself rather than take ours.

What we are not doing

We are not putting a valuation on this page. A pre-seed price is negotiated against the round size, the use of funds and the investor's own view of the market — not asserted by the founder in a data room. What this room commits to is that every input you would need to build your own view is here, sourced, and current: four years of CRM records, Xero actuals to 5 August 2026, the honest standalone win rate, the concentration, the negative working capital and its repair, and a commitment register that separates signed paper from conversation. Price it yourself. That is the point.

EBITDA proxy is net profit + depreciation ($11,877 + $17,684 = $29,561) for FY2026, unaudited. Multiple bands are published small-agency and data/analytics ranges cited on the Evidence tab, not a valuation of this company.

Pre-seed

Where the money goes: connection, not capex

The node fleet, vehicles, warehouse stock and staging are already on the balance sheet — roughly $105k of plant and equipment — and where a venue has incumbent cameras, Vision reads them without new hardware. A pre-seed here does not buy machines. It buys the connective tissue: the salesperson, the integrations, the benchmark, and the structures that let the data compound.

The physically digital premise

We are on site anyway — building the stage, running the show. The marginal cost of instrumenting an event we already deliver is hours, not capital: the client's site, the client's cameras where they exist, our nodes where they don't, and our system connecting all of it. That is why this budget is weighted to people and integration, not hardware.

Set the raise, shape the budget

Pre-seed raise

$500k

Deployment horizon

18 mo

Sales capacity — BD lead

30%

Product & integration engineering

25%

Benchmark deployment & site ops

15%

Design-partner program & GTM

10%

ANZ pathways — JV & CHQT establishment

Privacy, IP & compliance

Working capital & contingency

Monthly deployment

$28k/mo

raise ÷ horizon, evenly phased

Cost per benchmark site

$4k

benchmark allocation ÷ 20 sites

Weightings are relative — they always normalise to 100% of the raise.

What each dollar is for

Allocation	Share	Amount	What it buys
Sales capacity — BD lead	30%	$150k	12–18 months of a dedicated BD lead with events-sector relationships. The Horizon 1 constraint, and the first real payroll load.
Product & integration engineering	25%	$125k	Scaling what is already delivering — the vision layer against incumbent cameras, the LoRa mesh, the MQTT/telemetry pipeline — plus the agent layer build.
Benchmark deployment & site ops	15%	$75k	Instrumenting toward 20 comparable sites: calibration, install labour, node top-ups only where a site has no cameras.
Design-partner program & GTM	10%	$50k	Running the $3,995/event partner cohort, case studies, the renewal and indexation push on the anchors.
ANZ pathways — JV & CHQT establishment	8%	$40k	Moreton Bay ↔ Queenstown joint-venture and social-enterprise setup, CHQT launch support, international showcase travel.
Privacy, IP & compliance	7%	$35k	IP register and licence terms, privacy advice before any joined dataset is commercialised, contract templates.
Working capital & contingency	5%	$25k	Buffer against the negative working-capital starting point and event-cycle cash timing.
Total	100%	$500k

Illustration, not a forecast

Allocations are editable assumptions mapped to the Horizon 1–2 plan, not a costed budget. A term-sheet-grade budget needs the three-statement model (an open item on our side). Round mechanics — dilution, ESOP, SAFEs, the waterfall — live in the companion Capital Strategy & Cap Table model.

What the horizon has to produce

By	Funded by this budget	Proof it worked
+90 days	BD lead hired; graveyard reactivations called; analytics inclusion written into every production quote	Forward cover visibly moving toward $1.0m
Jan 2027	Design-partner cohort live and instrumented	$1.0m forward cover · paying design partners on the books
Dec 2027	Attached analytics converting to recurring site licences	>$100k ARR
Mid 2028	Benchmark instrumented across comparable sites	20+ sites · standalone re-entry unlocked

Investor prompts

Interrogate this yourself

Pick whichever assistant you already use. Every prompt below is self-contained — it carries the figures and context an outside model needs, so it works in a fresh chat with no prior knowledge of this room. Several are deliberately adversarial. We would rather you stress-test the thesis than take it on faith.

Which assistant are you using?

Gate 01 — ChatGPT flow

ChatGPT
Balanced flow — steady throughput, occasional bursts

ChatGPT. Turn web search on before running the market prompts — without it, it will answer the 2032 and RBA questions from stale training data and sound certain doing it. It is the strongest of these at holding a long adversarial brief without drifting into flattery.

On the numbers inside these prompts

Figures embedded below come from Artefact Group's own CRM and invoice register as at 31 July 2026. An outside model cannot verify them and may confidently invent corroboration — treat anything it adds beyond the pasted context as unverified, and ask us for the source file.

01 Brief me before the meeting›

A neutral primer. Use this first if you have not read the room — it gives your AI enough to hold a real conversation.

CONTEXT — Artefact Group Pty Ltd (Brisbane, Australia) and its product CrowdSense.
Artefact Group is a white-label events production company: it builds and runs events end-to-end (production, site and safety management, staging, hardware hire, warehousing, crew, content). CrowdSense is its crowd-analytics product — sensor nodes and camera-based computer vision producing live people counts, flow, occupancy, queues and heatmaps, plus a dashboard (Aura OS) and analytics layer (Pulse). The stated thesis is that running events generates labelled behavioural data as a by-product, so data acquisition cost is effectively negative, and that the resulting "collective brain" compounds with every event delivered.

VERIFIED FIGURES from the company's own CRM and invoice register as at 31 July 2026 (won contract value by close date, Australian financial years, AUD, unaudited):
- FY2023 $65,125 (11 deals) / FY2024 $223,785 (11) / FY2025 $345,750 (9) / FY2026 $527,259 (31)
- Approx. 100% CAGR on won value FY2023 to FY2026
- Win rate: 51% by deal count, 35.4% by value (63 won / 60 lost all time)
- Client concentration: top 3 accounts = 72.3% of FY2026; InStitchu = 33.7% of all revenue ever won
- Live pipeline as at 31 July 2026: 2 deals, $8,495. A further $241,500 sits dormant/"graveyard"
- Receivables reconciled against Xero as at 13 August 2026: $45,060 outstanding across 11 invoices / 5 debtors, 58.4% of it overdue ($26,327); the largest overdue line is a cycling-league client (ProVelo) at $12,531, largest total exposure is InStitchu at $23,598 across 4 invoices (most not yet due). An earlier register-derived figure of $129,073 was found on reconciliation to overstate receivables roughly 5x (register hygiene, not cash reality)
- Xero actuals FY2026 (accrual, unaudited): recognised revenue $446,002 and net profit $11,877 (FY2025: $301,369 and a $19,677 loss). As at 13 August 2026: cash at bank $66,963; payables $28,780 across 90 bills (98.5% overdue); working capital -$13,491; net equity $42,408. Recognised revenue differs from won contract value by timing: contracts count when won, revenue when earned
- CrowdSense sold standalone: 1 win ($4,500) against 10 losses ($169,500) — a 2.6% win rate by value. Lost bids include a regional tourism body ($75k), a sustainability certifier ($35k), a state tourism council ($25k) and two city councils
- CrowdSense delivered inside a production contract has worked: Brisbane Cycling Festival 2025 measured ~27,500 visits / ~22,500 unique attendees under the AusCycling production contract; AusCycling is now a $256,095 lifetime account

TASK: Brief me for a 45-minute first meeting with this company's founder. Produce:
1. A three-sentence plain-English summary of what the business actually does and where its money comes from today.
2. The strongest genuine argument for why this could become valuable.
3. The three things I should be most sceptical about, ranked.
4. Eight specific questions I should ask — each phrased so a vague answer is obvious.
Be concise and direct. Do not pad. If a figure above seems internally inconsistent, say so.

02 Red-team the whole thesis›

Deliberately adversarial. Ask your AI to argue the company should not be backed, and see whether the objections are survivable.

CONTEXT — Artefact Group Pty Ltd (Brisbane, Australia) and its product CrowdSense.
Artefact Group is a white-label events production company: it builds and runs events end-to-end (production, site and safety management, staging, hardware hire, warehousing, crew, content). CrowdSense is its crowd-analytics product — sensor nodes and camera-based computer vision producing live people counts, flow, occupancy, queues and heatmaps, plus a dashboard (Aura OS) and analytics layer (Pulse). The stated thesis is that running events generates labelled behavioural data as a by-product, so data acquisition cost is effectively negative, and that the resulting "collective brain" compounds with every event delivered.

VERIFIED FIGURES from the company's own CRM and invoice register as at 31 July 2026 (won contract value by close date, Australian financial years, AUD, unaudited):
- FY2023 $65,125 (11 deals) / FY2024 $223,785 (11) / FY2025 $345,750 (9) / FY2026 $527,259 (31)
- Approx. 100% CAGR on won value FY2023 to FY2026
- Win rate: 51% by deal count, 35.4% by value (63 won / 60 lost all time)
- Client concentration: top 3 accounts = 72.3% of FY2026; InStitchu = 33.7% of all revenue ever won
- Live pipeline as at 31 July 2026: 2 deals, $8,495. A further $241,500 sits dormant/"graveyard"
- Receivables reconciled against Xero as at 13 August 2026: $45,060 outstanding across 11 invoices / 5 debtors, 58.4% of it overdue ($26,327); the largest overdue line is a cycling-league client (ProVelo) at $12,531, largest total exposure is InStitchu at $23,598 across 4 invoices (most not yet due). An earlier register-derived figure of $129,073 was found on reconciliation to overstate receivables roughly 5x (register hygiene, not cash reality)
- Xero actuals FY2026 (accrual, unaudited): recognised revenue $446,002 and net profit $11,877 (FY2025: $301,369 and a $19,677 loss). As at 13 August 2026: cash at bank $66,963; payables $28,780 across 90 bills (98.5% overdue); working capital -$13,491; net equity $42,408. Recognised revenue differs from won contract value by timing: contracts count when won, revenue when earned
- CrowdSense sold standalone: 1 win ($4,500) against 10 losses ($169,500) — a 2.6% win rate by value. Lost bids include a regional tourism body ($75k), a sustainability certifier ($35k), a state tourism council ($25k) and two city councils
- CrowdSense delivered inside a production contract has worked: Brisbane Cycling Festival 2025 measured ~27,500 visits / ~22,500 unique attendees under the AusCycling production contract; AusCycling is now a $256,095 lifetime account

TASK: You are a sceptical investment committee member who is inclined to pass. Write the strongest possible case AGAINST backing this business. Attack the core claim that running events creates a defensible compounding data asset. Address specifically:
- Whether "negative-cost data acquisition" is a real moat or a re-description of being a services business
- Whether a 2.6% standalone win rate for the product means the product has no independent market
- Whether ~100% revenue CAGR off a $65k base is meaningful growth or just early-stage noise
- Whether $8,495 of live pipeline against a $527k year is a temporary gap or evidence of a structural sales failure
Then, having made the case against, state honestly what evidence would change your mind. Do not be gratuitously negative — be rigorous.

03 Is the bundling insight real, or an excuse?›

The single most important strategic claim in this room. Test whether it holds up.

CONTEXT — Artefact Group Pty Ltd (Brisbane, Australia) and its product CrowdSense.
Artefact Group is a white-label events production company: it builds and runs events end-to-end (production, site and safety management, staging, hardware hire, warehousing, crew, content). CrowdSense is its crowd-analytics product — sensor nodes and camera-based computer vision producing live people counts, flow, occupancy, queues and heatmaps, plus a dashboard (Aura OS) and analytics layer (Pulse). The stated thesis is that running events generates labelled behavioural data as a by-product, so data acquisition cost is effectively negative, and that the resulting "collective brain" compounds with every event delivered.

VERIFIED FIGURES from the company's own CRM and invoice register as at 31 July 2026 (won contract value by close date, Australian financial years, AUD, unaudited):
- FY2023 $65,125 (11 deals) / FY2024 $223,785 (11) / FY2025 $345,750 (9) / FY2026 $527,259 (31)
- Approx. 100% CAGR on won value FY2023 to FY2026
- Win rate: 51% by deal count, 35.4% by value (63 won / 60 lost all time)
- Client concentration: top 3 accounts = 72.3% of FY2026; InStitchu = 33.7% of all revenue ever won
- Live pipeline as at 31 July 2026: 2 deals, $8,495. A further $241,500 sits dormant/"graveyard"
- Receivables reconciled against Xero as at 13 August 2026: $45,060 outstanding across 11 invoices / 5 debtors, 58.4% of it overdue ($26,327); the largest overdue line is a cycling-league client (ProVelo) at $12,531, largest total exposure is InStitchu at $23,598 across 4 invoices (most not yet due). An earlier register-derived figure of $129,073 was found on reconciliation to overstate receivables roughly 5x (register hygiene, not cash reality)
- Xero actuals FY2026 (accrual, unaudited): recognised revenue $446,002 and net profit $11,877 (FY2025: $301,369 and a $19,677 loss). As at 13 August 2026: cash at bank $66,963; payables $28,780 across 90 bills (98.5% overdue); working capital -$13,491; net equity $42,408. Recognised revenue differs from won contract value by timing: contracts count when won, revenue when earned
- CrowdSense sold standalone: 1 win ($4,500) against 10 losses ($169,500) — a 2.6% win rate by value. Lost bids include a regional tourism body ($75k), a sustainability certifier ($35k), a state tourism council ($25k) and two city councils
- CrowdSense delivered inside a production contract has worked: Brisbane Cycling Festival 2025 measured ~27,500 visits / ~22,500 unique attendees under the AusCycling production contract; AusCycling is now a $256,095 lifetime account

The company has concluded from four years of records that CrowdSense fails when sold standalone (2.6% win rate by value, 10 losses to councils and tourism bodies) but succeeds when included inside a production contract. They plan to stop selling it standalone and instead include it in every delivery contract as standard, building a benchmark dataset before re-entering the standalone market.

TASK: Assess that conclusion critically.
1. Is this a genuine strategic insight, or a rationalisation of failing to sell a product?
2. What alternative explanations exist for the standalone losses — pricing, positioning, procurement process, buyer budget authority, product maturity, absence of a salesperson?
3. Name real companies that successfully made this services-to-product transition, and real ones that tried and stayed services businesses forever. What separated them?
4. What specific, measurable evidence in the next 18 months would prove this is working versus quietly failing?

04 Stress-test the market conditions›

Checks whether the tailwind narrative survives contact with the macro picture.

CONTEXT — Artefact Group Pty Ltd (Brisbane, Australia) and its product CrowdSense.
Artefact Group is a white-label events production company: it builds and runs events end-to-end (production, site and safety management, staging, hardware hire, warehousing, crew, content). CrowdSense is its crowd-analytics product — sensor nodes and camera-based computer vision producing live people counts, flow, occupancy, queues and heatmaps, plus a dashboard (Aura OS) and analytics layer (Pulse). The stated thesis is that running events generates labelled behavioural data as a by-product, so data acquisition cost is effectively negative, and that the resulting "collective brain" compounds with every event delivered.

VERIFIED FIGURES from the company's own CRM and invoice register as at 31 July 2026 (won contract value by close date, Australian financial years, AUD, unaudited):
- FY2023 $65,125 (11 deals) / FY2024 $223,785 (11) / FY2025 $345,750 (9) / FY2026 $527,259 (31)
- Approx. 100% CAGR on won value FY2023 to FY2026
- Win rate: 51% by deal count, 35.4% by value (63 won / 60 lost all time)
- Client concentration: top 3 accounts = 72.3% of FY2026; InStitchu = 33.7% of all revenue ever won
- Live pipeline as at 31 July 2026: 2 deals, $8,495. A further $241,500 sits dormant/"graveyard"
- Receivables reconciled against Xero as at 13 August 2026: $45,060 outstanding across 11 invoices / 5 debtors, 58.4% of it overdue ($26,327); the largest overdue line is a cycling-league client (ProVelo) at $12,531, largest total exposure is InStitchu at $23,598 across 4 invoices (most not yet due). An earlier register-derived figure of $129,073 was found on reconciliation to overstate receivables roughly 5x (register hygiene, not cash reality)
- Xero actuals FY2026 (accrual, unaudited): recognised revenue $446,002 and net profit $11,877 (FY2025: $301,369 and a $19,677 loss). As at 13 August 2026: cash at bank $66,963; payables $28,780 across 90 bills (98.5% overdue); working capital -$13,491; net equity $42,408. Recognised revenue differs from won contract value by timing: contracts count when won, revenue when earned
- CrowdSense sold standalone: 1 win ($4,500) against 10 losses ($169,500) — a 2.6% win rate by value. Lost bids include a regional tourism body ($75k), a sustainability certifier ($35k), a state tourism council ($25k) and two city councils
- CrowdSense delivered inside a production contract has worked: Brisbane Cycling Festival 2025 measured ~27,500 visits / ~22,500 unique attendees under the AusCycling production contract; AusCycling is now a $256,095 lifetime account

The company argues that South East Queensland is unusually favourable because of Brisbane 2032. Cited: A$7.1bn Queensland Games venue programme across 17 venues with construction commenced H1 2026; Tourism and Events Queensland grants rising 28% to A$96.25m in FY2026-27; Queensland visitor expenditure A$45.5bn for the year to March 2026 with international spend up 35.2%; Brisbane supporting 928 events and 2,036 performances in 2025.

Countervailing: RBA cash rate 4.35% after three increases in 2026 with markets pricing ~4.7% by year end; Australian headline inflation peaking ~4.8% in the June 2026 quarter; business investment growth forecast to fall to 0.8% by December 2026; Gartner reporting global marketing budgets up only 1.3% year on year at 7.8% of revenue; modern award wages up 4.75% from 1 July 2026; public liability insurance up 30-50% versus pre-COVID.

TASK: Which effect dominates for a small events production business over the next 24 months? Search for current data to verify or correct any of the figures above. Then state plainly whether the "good timing" claim survives, and what the company should do differently if it does not.

05 Sanity-check the market sizing›

The room deliberately declines to quote a TAM. Find out whether that was honesty or evasion.

CONTEXT — Artefact Group Pty Ltd (Brisbane, Australia) and its product CrowdSense.
Artefact Group is a white-label events production company: it builds and runs events end-to-end (production, site and safety management, staging, hardware hire, warehousing, crew, content). CrowdSense is its crowd-analytics product — sensor nodes and camera-based computer vision producing live people counts, flow, occupancy, queues and heatmaps, plus a dashboard (Aura OS) and analytics layer (Pulse). The stated thesis is that running events generates labelled behavioural data as a by-product, so data acquisition cost is effectively negative, and that the resulting "collective brain" compounds with every event delivered.

VERIFIED FIGURES from the company's own CRM and invoice register as at 31 July 2026 (won contract value by close date, Australian financial years, AUD, unaudited):
- FY2023 $65,125 (11 deals) / FY2024 $223,785 (11) / FY2025 $345,750 (9) / FY2026 $527,259 (31)
- Approx. 100% CAGR on won value FY2023 to FY2026
- Win rate: 51% by deal count, 35.4% by value (63 won / 60 lost all time)
- Client concentration: top 3 accounts = 72.3% of FY2026; InStitchu = 33.7% of all revenue ever won
- Live pipeline as at 31 July 2026: 2 deals, $8,495. A further $241,500 sits dormant/"graveyard"
- Receivables reconciled against Xero as at 13 August 2026: $45,060 outstanding across 11 invoices / 5 debtors, 58.4% of it overdue ($26,327); the largest overdue line is a cycling-league client (ProVelo) at $12,531, largest total exposure is InStitchu at $23,598 across 4 invoices (most not yet due). An earlier register-derived figure of $129,073 was found on reconciliation to overstate receivables roughly 5x (register hygiene, not cash reality)
- Xero actuals FY2026 (accrual, unaudited): recognised revenue $446,002 and net profit $11,877 (FY2025: $301,369 and a $19,677 loss). As at 13 August 2026: cash at bank $66,963; payables $28,780 across 90 bills (98.5% overdue); working capital -$13,491; net equity $42,408. Recognised revenue differs from won contract value by timing: contracts count when won, revenue when earned
- CrowdSense sold standalone: 1 win ($4,500) against 10 losses ($169,500) — a 2.6% win rate by value. Lost bids include a regional tourism body ($75k), a sustainability certifier ($35k), a state tourism council ($25k) and two city councils
- CrowdSense delivered inside a production contract has worked: Brisbane Cycling Festival 2025 measured ~27,500 visits / ~22,500 unique attendees under the AusCycling production contract; AusCycling is now a $256,095 lifetime account

The company deliberately publishes NO crowd-analytics market size, on the basis that three published 2026 estimates differ by up to 2.4x for the same year (US$4.84bn, US$1.99bn and US$1.64bn from different research houses), and that no credible Australian-specific figure exists in public sources.

TASK:
1. Search for current crowd analytics, people counting and venue intelligence market size estimates. Do the published figures actually conflict as described?
2. Is declining to quote a TAM a mark of intellectual honesty, or an evasion of a question the company should be able to answer?
3. Build a defensible BOTTOM-UP addressable market for crowd analytics at events and venues in South East Queensland — state every assumption, and show the arithmetic so I can challenge each input.
4. What would that number need to be for this to be a venture-scale opportunity rather than a good small business?

06 Competitive positioning›

Who else does this, and what happens when they notice?

CONTEXT — Artefact Group Pty Ltd (Brisbane, Australia) and its product CrowdSense.
Artefact Group is a white-label events production company: it builds and runs events end-to-end (production, site and safety management, staging, hardware hire, warehousing, crew, content). CrowdSense is its crowd-analytics product — sensor nodes and camera-based computer vision producing live people counts, flow, occupancy, queues and heatmaps, plus a dashboard (Aura OS) and analytics layer (Pulse). The stated thesis is that running events generates labelled behavioural data as a by-product, so data acquisition cost is effectively negative, and that the resulting "collective brain" compounds with every event delivered.

VERIFIED FIGURES from the company's own CRM and invoice register as at 31 July 2026 (won contract value by close date, Australian financial years, AUD, unaudited):
- FY2023 $65,125 (11 deals) / FY2024 $223,785 (11) / FY2025 $345,750 (9) / FY2026 $527,259 (31)
- Approx. 100% CAGR on won value FY2023 to FY2026
- Win rate: 51% by deal count, 35.4% by value (63 won / 60 lost all time)
- Client concentration: top 3 accounts = 72.3% of FY2026; InStitchu = 33.7% of all revenue ever won
- Live pipeline as at 31 July 2026: 2 deals, $8,495. A further $241,500 sits dormant/"graveyard"
- Receivables reconciled against Xero as at 13 August 2026: $45,060 outstanding across 11 invoices / 5 debtors, 58.4% of it overdue ($26,327); the largest overdue line is a cycling-league client (ProVelo) at $12,531, largest total exposure is InStitchu at $23,598 across 4 invoices (most not yet due). An earlier register-derived figure of $129,073 was found on reconciliation to overstate receivables roughly 5x (register hygiene, not cash reality)
- Xero actuals FY2026 (accrual, unaudited): recognised revenue $446,002 and net profit $11,877 (FY2025: $301,369 and a $19,677 loss). As at 13 August 2026: cash at bank $66,963; payables $28,780 across 90 bills (98.5% overdue); working capital -$13,491; net equity $42,408. Recognised revenue differs from won contract value by timing: contracts count when won, revenue when earned
- CrowdSense sold standalone: 1 win ($4,500) against 10 losses ($169,500) — a 2.6% win rate by value. Lost bids include a regional tourism body ($75k), a sustainability certifier ($35k), a state tourism council ($25k) and two city councils
- CrowdSense delivered inside a production contract has worked: Brisbane Cycling Festival 2025 measured ~27,500 visits / ~22,500 unique attendees under the AusCycling production contract; AusCycling is now a $256,095 lifetime account

TASK: Map the competitive landscape this company faces, across three distinct layers:
1. Crowd analytics / people counting vendors (for example Xovis, Density, Placer.ai, Vivacity, Crowd Connected, RetailNext, FootfallCam, V-Count, Milesight) — who could serve Australian events, and on what commercial model?
2. Australian and international events production and brand experience agencies operating in Queensland — including TBA Group, which entered Australia in November 2025 citing Brisbane 2032, and Encore, which acquired brand experience agency FIRST in December 2025 for its data and insights capability.
3. Venue and ticketing platforms that already hold attendance data and could extend into analytics.
For each layer: how real is the threat to a Brisbane operator of this size, what is the realistic defence, and what would a well-funded entrant have to do to make this company's position untenable?

07 Interrogate the unit economics›

The numbers the room does not publish, and why their absence matters.

CONTEXT — Artefact Group Pty Ltd (Brisbane, Australia) and its product CrowdSense.
Artefact Group is a white-label events production company: it builds and runs events end-to-end (production, site and safety management, staging, hardware hire, warehousing, crew, content). CrowdSense is its crowd-analytics product — sensor nodes and camera-based computer vision producing live people counts, flow, occupancy, queues and heatmaps, plus a dashboard (Aura OS) and analytics layer (Pulse). The stated thesis is that running events generates labelled behavioural data as a by-product, so data acquisition cost is effectively negative, and that the resulting "collective brain" compounds with every event delivered.

VERIFIED FIGURES from the company's own CRM and invoice register as at 31 July 2026 (won contract value by close date, Australian financial years, AUD, unaudited):
- FY2023 $65,125 (11 deals) / FY2024 $223,785 (11) / FY2025 $345,750 (9) / FY2026 $527,259 (31)
- Approx. 100% CAGR on won value FY2023 to FY2026
- Win rate: 51% by deal count, 35.4% by value (63 won / 60 lost all time)
- Client concentration: top 3 accounts = 72.3% of FY2026; InStitchu = 33.7% of all revenue ever won
- Live pipeline as at 31 July 2026: 2 deals, $8,495. A further $241,500 sits dormant/"graveyard"
- Receivables reconciled against Xero as at 13 August 2026: $45,060 outstanding across 11 invoices / 5 debtors, 58.4% of it overdue ($26,327); the largest overdue line is a cycling-league client (ProVelo) at $12,531, largest total exposure is InStitchu at $23,598 across 4 invoices (most not yet due). An earlier register-derived figure of $129,073 was found on reconciliation to overstate receivables roughly 5x (register hygiene, not cash reality)
- Xero actuals FY2026 (accrual, unaudited): recognised revenue $446,002 and net profit $11,877 (FY2025: $301,369 and a $19,677 loss). As at 13 August 2026: cash at bank $66,963; payables $28,780 across 90 bills (98.5% overdue); working capital -$13,491; net equity $42,408. Recognised revenue differs from won contract value by timing: contracts count when won, revenue when earned
- CrowdSense sold standalone: 1 win ($4,500) against 10 losses ($169,500) — a 2.6% win rate by value. Lost bids include a regional tourism body ($75k), a sustainability certifier ($35k), a state tourism council ($25k) and two city councils
- CrowdSense delivered inside a production contract has worked: Brisbane Cycling Festival 2025 measured ~27,500 visits / ~22,500 unique attendees under the AusCycling production contract; AusCycling is now a $256,095 lifetime account

Note that the company reports won CONTRACT VALUE from its CRM alongside Xero recognised revenue (FY2026: $446,002 income, $11,877 net profit; FY2025: $301,369, -$19,677 loss). Gross margin by job, a cash-flow model and a three-statement model are not yet published. Sensor hardware is described as 4MP edge-AI cameras with on-device inference, IP67 rated, with Wi-Fi, 4G and PoE options.

TASK:
1. What unit economics would I need to see before this is investable? Be specific about the metrics and the granularity.
2. For an events production business at roughly A$500k of annual contract value in Australia, what gross margin and EBITDA margin should I expect? What would be a red flag in either direction?
3. How should I think about the capital intensity of deploying sensor hardware across sites — payback period, utilisation, obsolescence, and who owns the asset?
4. What is the realistic cost to build and maintain the software layer described (live dashboard, analytics, computer vision across many camera brands) for a team of this size, and does that cost undermine the margin story?
5. Draft the exact schedule of financial information I should request before a second meeting.

08 Concentration and working capital›

The two risks most likely to kill this in the next twelve months.

CONTEXT — Artefact Group Pty Ltd (Brisbane, Australia) and its product CrowdSense.
Artefact Group is a white-label events production company: it builds and runs events end-to-end (production, site and safety management, staging, hardware hire, warehousing, crew, content). CrowdSense is its crowd-analytics product — sensor nodes and camera-based computer vision producing live people counts, flow, occupancy, queues and heatmaps, plus a dashboard (Aura OS) and analytics layer (Pulse). The stated thesis is that running events generates labelled behavioural data as a by-product, so data acquisition cost is effectively negative, and that the resulting "collective brain" compounds with every event delivered.

VERIFIED FIGURES from the company's own CRM and invoice register as at 31 July 2026 (won contract value by close date, Australian financial years, AUD, unaudited):
- FY2023 $65,125 (11 deals) / FY2024 $223,785 (11) / FY2025 $345,750 (9) / FY2026 $527,259 (31)
- Approx. 100% CAGR on won value FY2023 to FY2026
- Win rate: 51% by deal count, 35.4% by value (63 won / 60 lost all time)
- Client concentration: top 3 accounts = 72.3% of FY2026; InStitchu = 33.7% of all revenue ever won
- Live pipeline as at 31 July 2026: 2 deals, $8,495. A further $241,500 sits dormant/"graveyard"
- Receivables reconciled against Xero as at 13 August 2026: $45,060 outstanding across 11 invoices / 5 debtors, 58.4% of it overdue ($26,327); the largest overdue line is a cycling-league client (ProVelo) at $12,531, largest total exposure is InStitchu at $23,598 across 4 invoices (most not yet due). An earlier register-derived figure of $129,073 was found on reconciliation to overstate receivables roughly 5x (register hygiene, not cash reality)
- Xero actuals FY2026 (accrual, unaudited): recognised revenue $446,002 and net profit $11,877 (FY2025: $301,369 and a $19,677 loss). As at 13 August 2026: cash at bank $66,963; payables $28,780 across 90 bills (98.5% overdue); working capital -$13,491; net equity $42,408. Recognised revenue differs from won contract value by timing: contracts count when won, revenue when earned
- CrowdSense sold standalone: 1 win ($4,500) against 10 losses ($169,500) — a 2.6% win rate by value. Lost bids include a regional tourism body ($75k), a sustainability certifier ($35k), a state tourism council ($25k) and two city councils
- CrowdSense delivered inside a production contract has worked: Brisbane Cycling Festival 2025 measured ~27,500 visits / ~22,500 unique attendees under the AusCycling production contract; AusCycling is now a $256,095 lifetime account

TASK: Assess the two most immediate financial risks.
1. CONCENTRATION — top 3 clients are 72.3% of FY2026; the largest is 33.7% of all revenue ever. How should I price this risk? What contractual and structural mitigations actually work, and which are cosmetic? What discount would a trade acquirer apply?
2. WORKING CAPITAL — as at 13 August 2026, receivables are $45,060 (58.4% overdue), payables $28,780 (98.5% overdue), working capital -$13,491 and cash $66,963. Is this working-capital position dangerous or merely untidy for a business of this size, and what sequence of fixes would you require?
3. The company's CRM invoice register overstated receivables roughly 5x until it was reconciled against the accounting system in July 2026. What does that record-keeping failure itself tell me as an investor, and what remediation and reporting cadence would I require before closing?
4. Give me a short list of covenants or conditions I would reasonably require if I invested.

09 Technical due diligence›

For a technical investor or an advisor doing the engineering read.

CONTEXT — Artefact Group Pty Ltd (Brisbane, Australia) and its product CrowdSense.
Artefact Group is a white-label events production company: it builds and runs events end-to-end (production, site and safety management, staging, hardware hire, warehousing, crew, content). CrowdSense is its crowd-analytics product — sensor nodes and camera-based computer vision producing live people counts, flow, occupancy, queues and heatmaps, plus a dashboard (Aura OS) and analytics layer (Pulse). The stated thesis is that running events generates labelled behavioural data as a by-product, so data acquisition cost is effectively negative, and that the resulting "collective brain" compounds with every event delivered.

VERIFIED FIGURES from the company's own CRM and invoice register as at 31 July 2026 (won contract value by close date, Australian financial years, AUD, unaudited):
- FY2023 $65,125 (11 deals) / FY2024 $223,785 (11) / FY2025 $345,750 (9) / FY2026 $527,259 (31)
- Approx. 100% CAGR on won value FY2023 to FY2026
- Win rate: 51% by deal count, 35.4% by value (63 won / 60 lost all time)
- Client concentration: top 3 accounts = 72.3% of FY2026; InStitchu = 33.7% of all revenue ever won
- Live pipeline as at 31 July 2026: 2 deals, $8,495. A further $241,500 sits dormant/"graveyard"
- Receivables reconciled against Xero as at 13 August 2026: $45,060 outstanding across 11 invoices / 5 debtors, 58.4% of it overdue ($26,327); the largest overdue line is a cycling-league client (ProVelo) at $12,531, largest total exposure is InStitchu at $23,598 across 4 invoices (most not yet due). An earlier register-derived figure of $129,073 was found on reconciliation to overstate receivables roughly 5x (register hygiene, not cash reality)
- Xero actuals FY2026 (accrual, unaudited): recognised revenue $446,002 and net profit $11,877 (FY2025: $301,369 and a $19,677 loss). As at 13 August 2026: cash at bank $66,963; payables $28,780 across 90 bills (98.5% overdue); working capital -$13,491; net equity $42,408. Recognised revenue differs from won contract value by timing: contracts count when won, revenue when earned
- CrowdSense sold standalone: 1 win ($4,500) against 10 losses ($169,500) — a 2.6% win rate by value. Lost bids include a regional tourism body ($75k), a sustainability certifier ($35k), a state tourism council ($25k) and two city councils
- CrowdSense delivered inside a production contract has worked: Brisbane Cycling Festival 2025 measured ~27,500 visits / ~22,500 unique attendees under the AusCycling production contract; AusCycling is now a $256,095 lifetime account

The technology stack is described as: edge-AI counting cameras (4MP, 2688x1520, ~0.6 TOPS on-device inference, up to 25fps, IP67, -20C to 50C, Wi-Fi 6 / LTE Cat.1 / Ethernet, interchangeable 51/88/137 degree lenses); crowd pressure sensors; a camera-brand-agnostic computer vision layer (3,600+ brand stack) described by the company as in delivery and expanding; LoRa mesh networking (Meshtastic, ESP32-based PAX counting nodes) for off-grid multi-node sites, camera zone-handoff logic, and an MQTT + Grafana telemetry pipeline — all described as in delivery mode and growing; a live dashboard and an analytics/forecasting layer.

TASK: Conduct a technical due diligence assessment.
1. Is this architecture credible for outdoor, temporary, often off-grid event deployments? What are the specific failure modes — power, backhaul, weather, occlusion, calibration drift, crowd density saturation?
2. How much of this is genuinely proprietary versus integration of commodity hardware and open-source components? Where would real defensible IP sit in a stack like this?
3. What accuracy should be expected for people counting in dense crowds, and how would I independently verify a vendor's claim?
4. What are the privacy, consent and regulatory implications of camera-based crowd analytics in Australia under the Privacy Act — and how does that differ if movement data is joined to ticketing identity?
5. What would you want to see demonstrated live, on site, before believing the capability?

10 Triangulate the valuation›

What is it worth, and which multiple applies?

CONTEXT — Artefact Group Pty Ltd (Brisbane, Australia) and its product CrowdSense.
Artefact Group is a white-label events production company: it builds and runs events end-to-end (production, site and safety management, staging, hardware hire, warehousing, crew, content). CrowdSense is its crowd-analytics product — sensor nodes and camera-based computer vision producing live people counts, flow, occupancy, queues and heatmaps, plus a dashboard (Aura OS) and analytics layer (Pulse). The stated thesis is that running events generates labelled behavioural data as a by-product, so data acquisition cost is effectively negative, and that the resulting "collective brain" compounds with every event delivered.

VERIFIED FIGURES from the company's own CRM and invoice register as at 31 July 2026 (won contract value by close date, Australian financial years, AUD, unaudited):
- FY2023 $65,125 (11 deals) / FY2024 $223,785 (11) / FY2025 $345,750 (9) / FY2026 $527,259 (31)
- Approx. 100% CAGR on won value FY2023 to FY2026
- Win rate: 51% by deal count, 35.4% by value (63 won / 60 lost all time)
- Client concentration: top 3 accounts = 72.3% of FY2026; InStitchu = 33.7% of all revenue ever won
- Live pipeline as at 31 July 2026: 2 deals, $8,495. A further $241,500 sits dormant/"graveyard"
- Receivables reconciled against Xero as at 13 August 2026: $45,060 outstanding across 11 invoices / 5 debtors, 58.4% of it overdue ($26,327); the largest overdue line is a cycling-league client (ProVelo) at $12,531, largest total exposure is InStitchu at $23,598 across 4 invoices (most not yet due). An earlier register-derived figure of $129,073 was found on reconciliation to overstate receivables roughly 5x (register hygiene, not cash reality)
- Xero actuals FY2026 (accrual, unaudited): recognised revenue $446,002 and net profit $11,877 (FY2025: $301,369 and a $19,677 loss). As at 13 August 2026: cash at bank $66,963; payables $28,780 across 90 bills (98.5% overdue); working capital -$13,491; net equity $42,408. Recognised revenue differs from won contract value by timing: contracts count when won, revenue when earned
- CrowdSense sold standalone: 1 win ($4,500) against 10 losses ($169,500) — a 2.6% win rate by value. Lost bids include a regional tourism body ($75k), a sustainability certifier ($35k), a state tourism council ($25k) and two city councils
- CrowdSense delivered inside a production contract has worked: Brisbane Cycling Festival 2025 measured ~27,500 visits / ~22,500 unique attendees under the AusCycling production contract; AusCycling is now a $256,095 lifetime account

Relevant market context: published Australian and global agency M&A multiples put small agencies under US$1m EBITDA at roughly 3-5x EBITDA, mid-market at 5-8x, and larger or high-performing data and analytics focused agencies at 8-12x. Acquirers reportedly pay premiums for recurring revenue above 80%, EBITDA margins above 20%, client retention above 90% and low owner dependency. Recent event-technology M&A has been explicitly data-driven — Cvent acquired ON24 (US$400m) and Goldcast (~US$300m) in December 2025; Bending Spoons agreed to acquire Eventbrite for ~US$500m; Encore acquired FIRST citing its data and insights capability.

TASK:
1. Which multiple realistically applies to this company TODAY, given its actual characteristics? Justify it against each of the four premium criteria.
2. What would it need to change, and over what period, to move up a tier — and is that change plausible for a business of this size and team?
3. Value the business today under three scenarios (conservative, base, optimistic), stating every assumption.
4. Is the gap between the current multiple and the target multiple worth more or less than simply growing revenue at the current rate without changing the revenue mix? Show the arithmetic.
5. What is the most likely realistic exit for a business like this in Australia, and who are the plausible acquirers by name?

11 What should I ask that I have not thought of?›

A closing sweep for anything the framing of this room might have led you past.

CONTEXT — Artefact Group Pty Ltd (Brisbane, Australia) and its product CrowdSense.
Artefact Group is a white-label events production company: it builds and runs events end-to-end (production, site and safety management, staging, hardware hire, warehousing, crew, content). CrowdSense is its crowd-analytics product — sensor nodes and camera-based computer vision producing live people counts, flow, occupancy, queues and heatmaps, plus a dashboard (Aura OS) and analytics layer (Pulse). The stated thesis is that running events generates labelled behavioural data as a by-product, so data acquisition cost is effectively negative, and that the resulting "collective brain" compounds with every event delivered.

VERIFIED FIGURES from the company's own CRM and invoice register as at 31 July 2026 (won contract value by close date, Australian financial years, AUD, unaudited):
- FY2023 $65,125 (11 deals) / FY2024 $223,785 (11) / FY2025 $345,750 (9) / FY2026 $527,259 (31)
- Approx. 100% CAGR on won value FY2023 to FY2026
- Win rate: 51% by deal count, 35.4% by value (63 won / 60 lost all time)
- Client concentration: top 3 accounts = 72.3% of FY2026; InStitchu = 33.7% of all revenue ever won
- Live pipeline as at 31 July 2026: 2 deals, $8,495. A further $241,500 sits dormant/"graveyard"
- Receivables reconciled against Xero as at 13 August 2026: $45,060 outstanding across 11 invoices / 5 debtors, 58.4% of it overdue ($26,327); the largest overdue line is a cycling-league client (ProVelo) at $12,531, largest total exposure is InStitchu at $23,598 across 4 invoices (most not yet due). An earlier register-derived figure of $129,073 was found on reconciliation to overstate receivables roughly 5x (register hygiene, not cash reality)
- Xero actuals FY2026 (accrual, unaudited): recognised revenue $446,002 and net profit $11,877 (FY2025: $301,369 and a $19,677 loss). As at 13 August 2026: cash at bank $66,963; payables $28,780 across 90 bills (98.5% overdue); working capital -$13,491; net equity $42,408. Recognised revenue differs from won contract value by timing: contracts count when won, revenue when earned
- CrowdSense sold standalone: 1 win ($4,500) against 10 losses ($169,500) — a 2.6% win rate by value. Lost bids include a regional tourism body ($75k), a sustainability certifier ($35k), a state tourism council ($25k) and two city councils
- CrowdSense delivered inside a production contract has worked: Brisbane Cycling Festival 2025 measured ~27,500 visits / ~22,500 unique attendees under the AusCycling production contract; AusCycling is now a $256,095 lifetime account

TASK: I have been given a well-organised investor room by this company. It is transparent about several weaknesses — an $8,495 live pipeline, 72.3% client concentration, $45,060 of receivables of which 58.4% is overdue, and a 2.6% standalone win rate for its product.

A well-presented set of disclosed weaknesses can itself be a framing device: it builds trust and directs attention toward the problems the company has already decided how to answer.
1. What important questions does this framing lead me AWAY from?
2. What would a company in this position most plausibly omit entirely rather than disclose and mitigate?
3. What should I independently verify rather than accept, and specifically how would I verify each item?
4. Give me five questions that would be genuinely uncomfortable to answer, and explain what a good answer to each would sound like versus an evasive one.

Risk register

What could stop this working

Ordered by how likely they are to bite in the next 12 months. Nothing here is hedged — an investor will find all of it in diligence, and finding it here first is worth more than the discomfort.

1 · Pipeline exhaustionCriticalNow

$8,495 of live pipeline against a $527k trailing year. There is no scenario in which the rest of this strategy matters if this is not fixed first. Event sales cycles run months, so the gap is already baked into the first half of FY2027.

Mitigation — Horizon 1 exists solely for this. $271,827 of warm, identified value is reachable without new product or market. The binding constraint is that no one currently owns selling as a full-time job.

2 · Client concentrationCriticalNow

Top three accounts are 72.3% of FY2026. InStitchu alone is 33.7% of all revenue ever won. Losing any one of the three is a business-threatening event, and all three renew annually.

Mitigation — Multi-year indexed terms before the counterparty tests the market; second and third anchor sectors in Horizon 2; concentration below 20% as an explicit 2029 gate.

3 · Working capital locked in receivablesCriticalNow

Working capital is negative — −$13,491 as at 13 August 2026 (was −$26,511 at 31 July). Receivables have grown to $45,060, of which 58.4% ($26,327) is overdue — $17,410 of that three-plus months — while payables of $28,780 are 98.5% overdue, and $13,271 of that is owed to a key delivery partner. At a 4.35% cash rate with business investment growth falling to 0.8%, external working capital is expensive and getting more so.

Mitigation — Reconciliation is done (31 Jul 2026 — the register overstated AR ~5×). Now: enforce terms on the ProVelo and Check 1-2 lines, make deposit-and-progress billing standard, and clear the aged payables before any diligence process opens the books.

4 · Key person dependencyHighStructural

Relationships, technical architecture, commercial judgement and delivery oversight all concentrate in the founder. Acquirers explicitly discount for this, and it is one of the four criteria separating a 3–5× multiple from an 8–12× one.

Mitigation — The agent layer is itself the mitigation: encoding operating knowledge into a system is what makes it transferable. Plus the first BD hire in Horizon 1 and documented delivery runbooks.

5 · CrowdSense never becomes a productHighH2

The plan assumes attached analytics can be converted to recurring licences once a benchmark exists. If clients keep treating measurement as a nice-to-have inside a production fee, the multiple never re-rates and this stays a services business with good tooling.

Mitigation — Treat the 20-site benchmark as the go/no-go gate. If it is live and licences still do not sell by mid-2028, stop investing in productisation and optimise the services business honestly instead.

6 · Better-capitalised entrantsHigh2026–28

TBA Group entered Australia in Nov 2025 explicitly citing Brisbane 2032. Encore acquired FIRST for its data and insights capability. These groups can buy the local position we intend to build, and they do not need it to be profitable early.

Mitigation — Compete on what capital cannot buy quickly: existing SEQ operator relationships, and longitudinal local data that only accrues with time on site. Speed to the benchmark matters more than breadth.

7 · Margin compressionMediumNow

Award wages +4.75% from 1 July 2026, public liability insurance +30–50% versus pre-COVID, against client marketing budgets growing 1.3%. A labour-and-logistics business absorbs all of that.

Mitigation — Contractual indexation at every renewal; shift mix toward measurement and licence revenue that does not scale with crew hours; warehouse hire to improve asset utilisation.

8 · Privacy and consent exposureMediumStructural

The strategy's value depends on joining movement data to identity. Under the Privacy Act, and given camera-based sensing is drawing regulatory attention internationally, a consent or provenance failure would not just be a fine — it would devalue the core asset.

Mitigation — Consent and provenance recorded at collection, not retrofitted; prefer anonymous counting where identity adds no analytical value; make privacy provenance an explicit acquisition screen. Take formal privacy advice before commercialising joined datasets.

9 · Data asset is asserted, not provenMediumNow

The compounding claim is the centre of the equity story, but the deployed base today is small. A sophisticated investor will ask how many sites, how many event-days, and what the benchmark actually contains — and the honest answer is currently modest.

Mitigation — Publish the real deployment count and event-day total rather than implying scale. Growth in that number is the most credible progress metric available and should be reported every quarter.

10 · Games timing riskLow–Medium2028–32

The 2032 thesis assumes SEQ event volume and measurement requirements rise on schedule. Venue programmes slip, and the official $8.1bn Queensland benefit estimate is from June 2021 and states no formal cost-benefit analysis was undertaken.

Mitigation — Do not underwrite anything on Games volume. Treat 2032 as upside on a business that must work on today's calendar of 928 Brisbane events and 853 BCEC events a year.