HEARO Goes Live On Both App Stores, FUN SQUEEZER On Google Play

Published on: Mon, Oct. 19, 2026 (SGT) · Status and modelled financials for all 21 Capital Vision projects, re-verified the same day · Get the next report

Evidence labels: ACTUAL measured fact as of 19 Oct 2026 · MODELLED output of a planning model · TARGET a goal, not a result. Models stand alone and are not added together.

1. HEARO: both stores live, now the brands have to pay

HEARO turns one photo and one sentence into a complaint that is classified, written up and routed to the brand's decision-maker. Identical complaints are grouped into a rally, so a brand hears from a group rather than one person. Consumers never pay. Revenue is designed to come from brands, through the HEARO Brand Response Desk (S$390 a month, MODELLED) and Defect Intelligence Reports (S$1,200 each, MODELLED).

What is live ACTUAL

Capital model M12, re-based on 19 Oct MODELLED

The model changed for two reasons, both measured. First, R&D completion rose from 45% to 67%: four of six work packages are now delivered (AI engine, interface, rally clustering, native apps on both stores). Second, the AI cost per case is now taken from live usage, S$0.03 instead of S$0.04. The remaining capital goes to the Brand Desk console, the moderation console, trademarks and patent, launch seeding, the first brand pilots and runway.

Stress test. The base case assumes 20 paying brands (TARGET). At 8 brands, annualised ROI falls to 35%, payback stretches to about 39 months and 36-month MOIC is 0.91x. HEARO needs at least 9 paying brands to return 1.0x within 36 months.

Next proof: the first paying Brand Response Desk, and the first rally of 10 people delivered together to a brand. · Open the HEARO capital account · Brand page

2. FUN SQUEEZER: S$1 of ads, S$1.63 back

FUN SQUEEZER is built for students. They photograph a homework page, and the app turns it into one card per question. Each card opens with the single idea the teacher is testing, delivered by a confident future version of the student. There is no subscription. Students buy credits and pay per question, so every paying user produces cash from the first question.

What is live ACTUAL

Ad spend versus future cash flow MODELLED

How one month of advertising pays back. In the scale phase, S$6,000 of ads brings 9,000 installs and 450 new paying users. Over their lifetime those users contribute S$9,756, which is S$1.63 for every S$1 spent. Each monthly budget recovers its cost in about four months, then keeps paying. Stacked month after month, cumulative contribution overtakes cumulative ad spend in month 12. By month 36 the model holds about 3,000 paying app users and 300 web users.

Base-case assumptions: 5% of installs become paying users and every paid install brings 2 organic installs (TARGET); S$2 per install, 15% monthly churn and 80 questions per paying user a month (MODELLED). The S$6,000 monthly scale budget (gate G4) is released only after the paid test confirms conversion, and a cost per install within the S$3.25 the unit economics can carry.

Next proof: App Store approval, the first production in-app purchase, and the question counts of the first 30 paying users. · Open the FUN SQUEEZER capital account · Try it on the web

3. All projects: status on 19 Oct 2026

Every project was re-checked against the live store and our event ledger today. "This week" comes from the ledger: Material change means a recorded event in the last 7 days. Awaiting evidence means no capital event has been recorded yet.

Model figures: capital envelope · annualised ROI · payback in months. ROI n/a means the model does not yet produce a steady-state return. Red marks a payback longer than 36 months on current inputs.

4. Actuarial review

Today we recomputed all 16 capital models: 15 live engines plus the CARE EASE brand-loop model. Each model answers one question, how much capital takes that project to its next stage. Models are never added together.

• Re-based this cycle: M12 HEARO, after store launch and measured AI cost. Updated: M10 FUN SQUEEZER now reflects live store status.
• Ad-gated growth: FUN SQUEEZER returns S$1.63 per ad dollar in its base case and turns cash positive in month 24. Its scale budget is released only on measured conversion, so capital follows proof.
• Long payback: Taiwan Rice (74.8 months) and MASOGIA (38.5 months) need a price, volume or channel change before they warrant outside capital.
• Return not yet computable: M02 SIGHTREAD platform and M04 ATITUDE produce a payback month but not a steady-state ROI.
• Largest envelopes are system builds: M01 unmanned stores (S$7.6M), M11 Utility Core (S$3.6M), M04 ATITUDE (S$3.0M) and M02 SIGHTREAD platform (S$2.6M). These are sized plans, not commitments, and none has a physical unit yet.
• Evidence depth is still building. Measured inputs are a minority in every model, for example 8 of 58 in M01, 11 of 40 in M10 and 2 of 17 in M12. Both apps are pre-revenue, so their revenue lines stay TARGET until the first payments arrive.
• Five retail brand lines are on the shelf or in the catalogue but have no capital model. Landed cost and sell-through are missing. The next step is to feed sales velocity from our store system into Capital Vision.

5. What changed this week (event ledger)

1. EV-HRO-0003 · HEARO passed gate G1: native apps live on App Store and Google Play. ACTUAL
2. EV-HRO-0004 · HEARO model M12 re-based to v1.1 (S$70,876 → S$60,848). MODELLED
3. EV-FSQ-0004 · FUN SQUEEZER live on Google Play; App Store in expedited review. ACTUAL
4. EV-FSQ-0005 · FUN SQUEEZER model M10 updated to v2.1. MODELLED
5. EV-VTFR-0001 · Vital Fruit first physical stock, 4 SKUs and 144 units. ACTUAL
6. EV-KKUN-0001 · EV-JHOM-0001 · Retail stock re-verified: King Kung 0, Jia Her Home 0. ACTUAL

6. Get the next report

7. Where to go next

Open Capital Vision »Get HEAROGet FUN SQUEEZER

Every number above can be adjusted on the project's capital account page, and the model recomputes as you change it. To discuss a project, reply to the report email or write to arthur@authentictaiwan.com.