EDGE AI • PRIVACY-FIRST

AI-Powered Childcare Safety Monitoring

Computer vision that helps CCTV spot safety incidents as they happen

Investor Deck | May 2026

1st Place, Healthcare Track — AI Singapore Pan-SEA AI Developer Challenge 2025

99%

of CCTV footage
is never reviewed

99% of CCTV Footage is Dark Data

CCTV records 24/7, but most footage is checked only after something happens
Enterprise systems can run $100K+, out of reach for most childcare centres
Cloud systems can send children's faces to third-party servers, creating PDPA risk

On-Premise Edge AI + Cloud Dashboard

Works with the cameras centres already have

ON-PREMISENO FACE UPLOADS

AI Incident Detection — Flags falls, rough handling, and altercations in real time
Privacy-First — Video stays on-site. Built for PDPA compliance.
Secure Cloud Dashboard — Detection records, auto-reports, and natural-language video search

CCTV
Existing cameras

Edge AI
On-premise box

Dashboard
Cloud reports

Alert
Instant notify

$1.8–5.2B Global TAM

Childcare safety AI monitoring

Total Addressable Market

AI Video Surveillance $3.9–6.5B

21–31% CAGR

Serviceable Addressable Market

ASEAN Pricing SAM ~$94M/yr

BASE CASE; $70–201M RANGE

Serviceable Obtainable Market

Y3 SOM $1.7M revenue

Singapore → Malaysia → ASEAN

AI Video Surveillance

$3.9–6.5B

21–31% CAGR by 2030

Childcare-Specific AI

$1.8–5.2B

Our Niche

School Security

$1.9–3.0B

11–21% CAGR by 2030

~5–10% of total AI surveillance by 2030

~21,000+ Near-Term Addressable Centres

🇸🇬

Singapore

~1,800

CCTV MANDATORY

🇲🇾

Malaysia

~3,400

Proposed

🇮🇩

Indonesia

~11,900

Long-term

🇹🇭

Thailand

~1,000

Long-term

🇵🇭

Philippines

~3,150

Long-term

REGULATORY BEACHHEAD → ASEAN SCALE

SAM: ~$94M/year base

~$36M/year recurring AMC beachhead

Blue Ocean — Nobody Owns This Niche

	Pricing	Deployment	Childcare AI	COGS
SIMIS.AI	Per-site ✓	On-premise ✓	Childcare AI ✓	<$500
Verkada	Per-camera ✗	Cloud ✗	Generic ✗	$600–3,500
Rhombus	Per-camera ✗	Cloud ✗	Generic ✗	$200–1,648
WatchMeGrow	Per-site ✓	Cloud ✗	Streaming only ✗	Custom

SIMIS 3-year TCO

Cloud competitors

80–90% lower 3-year TCO than Verkada

Real-Time AI Detection Engine

Model validation with bounding boxes and confidence scoring

AI model validation grid showing bounding boxes and confidence scores

Multi-class detection — Falls, rough handling, altercations, crowd density
Real-time inference — Processes at 15+ FPS on edge hardware
Confidence scoring — Filters false positives before alerts
Continuous learning — Fine-tuned on childcare-specific footage

15+ FPSBOUNDING BOXESEDGE FILTERING

Data Flywheel Effect

Every deployed centre generates labelled data → improves model → attracts more centres

Three Revenue Models

Premium — SG

$1,500 + $250/mo

Full-featured premium tier for Singapore's regulated market

3yr LTV

$10,500

Anchor offer — highest margin

Standard — MY

$800 + $180/mo

Mid-tier for Malaysia — competitive pricing with strong value

3yr LTV

$7,280

Regional expansion

Budget — ASEAN

$500 + $120/mo

Affordable entry tier for emerging ASEAN markets

3yr LTV

$4,820

Mass-market volume play

70–84% install · ~95% AMC | Software = 78–88% of revenue

3-Year Revenue Roadmap

Year 1

$240K

50 centres

Investment phase

Year 2

$645K

150 centres

Unit profitable

Year 3

$1.7M

400 centres

54% EBITDA

3-Year Cumulative: ~$2.6M | Company break-even: ~100–120 centres (Year 2)

Real Numbers, No Surprises

	Year 1	Year 2	Year 3
Revenue	$240K	$645K	$1,700K
COGS	$19.5K	$43.4K	$96.2K
Gross Profit	$220.5K	$601.6K	$1,603.8K
Gross Margin	91.9%	93.3%	94.3%
Operating Expenses	$256K	$352K	$679K
EBITDA	–$35.5K	+$249.6K	+$924.8K

92–95%Gross margin

Y2Unit profitable

$925KY3 EBITDA

Product gross margin: 92–95%. Losses are R&D investment, not business model.

Current Reality Tree

TOC — What to Change? 3 INTERVIEWS → 10 UDEs

Core constraint: CCTV captures everything, understands nothing

Safety incidents discovered too late — after harm has occurred

Parent complaints & trust erosion

Staff accountability gaps

ECDA/MSF compliance risk

Incidents found avg 2–3 days later — only through manual review or parent reports

Generic CCTV analytics too noisy; cloud AI blocked by PDPA concerns

CCTV = passive recording, no active intelligence layer

Framework: Theory of Constraints (Goldratt) — 3 operator interviews confirmed all UDEs

Future Reality Tree

TOC — What to Change To?

Injection: On-premise edge AI turns passive CCTV into active safety intelligence

Proactive safety monitoring replaces reactive incident discovery

Incidents caught in minutes

Response: hours → minutes

Accountability via objective records

Records: manual → auto time-stamped

Franchisor multi-site view

Oversight: site-by-site → centralised

Parents trust the system

Trust: complaints → confidence

Real-time CV detection + instant alerts

Auto incident reports + searchable footage

PDPA compliance by design

On-premise edge AI device connects to existing CCTV

Childcare-specific CV models + privacy-first architecture

Before

Reactive

After

Proactive

Framework: Theory of Constraints (Goldratt)

Negative branches pruned: teacher surveillance concern, alert fatigue, hardware cost barrier, parent livestream demand

Seeking $150K–300K for 12 Months

M1–4

$18–29K

Working prototype + dashboard MVP

M5–8

$19–49K

Pilot deployments + model fine-tuning

M9–12

$40–136K

50 units, certification, and first customers

Milestone 8: Go / no-go checkpoint

Grant Strategy

Startup SG Tech ($50K) + NRF AI SG ($250K) + Angel ($100K) = $400K

Why SIMIS.AI Wins

Supply Chain

Shenzhen co-founder gives us a hardware cost advantage

Data Flywheel

Each centre adds labelled examples, making the model harder to copy

Regulatory First-Mover

Singapore's childcare CCTV mandate gives us a clear first market

Hardware Lifecycle

Circular economy — hardware refresh and trade-in creates recurring revenue lock-in

$150K–300K funds 12 months to prove the wedge in a ~$94M pricing SAM

SIMIS.AI — A Trained Pair of Eyes for Every Camera