Who pays for the device — retailer or shopper?

You pay for the value you get — never the full device price. Options include a one-time entry fee, a monthly subscription, or loyalty vouchers. Category partners and the retailer co-fund the rest through retail media.

What data leaves the device? Is it GDPR/UK-compliant? Who owns it?

Only encrypted event metadata leaves the device; raw images are not stored or transmitted. Data is encrypted in transit and at rest, hosted on local cloud servers, and owned by the retailer under agreed access controls.

Where do the numbers come from?

Every figure is a sourced fact, a derived estimate (explained), or a benchmark with disclaimer. UK figures draw on ONS and DEFRA Family Food estimates; uplift assumptions are pilot projections.

SMART SHELF AI

Smart Shelf AI

The home fridge is your new sales channel

A retrofit inventory + auto-replenishment platform that turns every household’s breakfast-fridge consumption into your incremental revenue — with the in-home consumption signal you've never had.

Works with any fridge90%+ SKU recognitionUK data residencyPatent pending · EP25203165.3

Every forgotten item is a sale that leaks away

Your customer opens the fridge — out of milk. They were just at your store and didn’t remember. The trip is wasted, the basket is smaller, and the next purchase may happen at your competitor.

28.6M

UK households — addressable

ONS, 2024

~£570 / yr

breakfast-fridge spend per household

~ · ONS + DEFRA Family Food estimates

~£16bn

of breakfast-fridge staples bought for the home each year

~ · ~£570 × 28.6M households

+20%

incremental basket/re-order uplift we unlock

pilot projection

£470 / yr

breakfast-fridge waste we help cut

WRAP (household benchmark)

From shelf to re-order, automatically

Sense → Sync → Recognise → Re-order. The device captures consumption, the cloud turns it into SKU-level signal, and your e-commerce cart or subscription fires the re-order.

Sense

Weight sensors detect consumption — even through sealed packs and varying fridge zones.

Sync to UK cloud

Encrypted, cropped product-zone frames and weight events sync to UK-based servers. The camera only ever faces the product zone — faces and rooms are physically out of view.

Recognise

Cloud AI fuses cropped vision + weight for 90%+ SKU accuracy — your brands vs competitors.

Re-order

Low-stock triggers flow into your e-commerce cart or subscription via REST API.

Integration · REST API · 4–6 weeks to integrate · sandbox available

What this does to your numbers

UK · breakfast-fridge model

+20%

incremental basket / re-order uplift

~£114 / yr

incremental GMV per active household

−30%

less in-home breakfast-fridge waste

~£1.14M

incremental GMV / year per 10,000 active devices · UK

UK model: ~£570 breakfast-fridge spend/household × +20% uplift on 10,000 households. Pilot benchmarks; methodology available on request.

You pay for the value you get — never the full £199

Three ways to get the device

Category-backed

Recommended

£49 one-time · category £120 · retailer £30

Subscription

£5/month · £0 upfront

Loyalty vouchers

Pay with vouchers · no cash out of pocket

The retailer recovers its share through retail media, not grocery markup. Funded by category partners, not the shopper.

£199 vs. a £2,400 smart fridge

Why retrofit wins

FeatureSmart Shelf PlatformSmart Fridge

Entry price£49 one-time · £5/mo · loyalty vouchers£2,400–3,300 appliance

Install5-minute installDelivery + install + replacement

ReachWorks with 28M+ UK fridges<5% penetration · 10–15 yr cycle

CameraInward (privacy)Often faces kitchen

OwnershipYour brand, your APITheir ecosystem, their data

12–15× cheaper for your customers — and instantly deployable at scale.

Technology & privacy by design

Built for retail buyers and CISOs

UK data residency

Recognition and data hosted on UK-based cloud servers.

Encrypted & access-controlled

Data encrypted in transit and at rest; the retailer owns it.

Zone control

Items outside the tracked zone are simply ignored.

Camera faces the product

By design, never sees users, rooms or faces.

Under the hood

Built on the Realtek RTL8735BDM with a 2 MP / 179° camera, four precision weight sensors and a temperature / humidity sensor. The 20,000 mAh Li-ion battery is rated for normal fridge conditions (+4 °C) and delivers ~6 months of operation with a recharge cycle every 6 months. The camera faces only the device zone — full-room frames are never captured. On-device processing crops and obfuscates everything outside the product zone; only encrypted, cropped product-zone frames and weight events are transmitted to UK-based cloud servers for recognition. Cloud-processed images are not retained after recognition; only SKU-level event metadata is stored under the retailer’s agreed retention policy. Data is owned by the retailer under agreed controller/processor terms. A 'dead' device is surfaced via the dashboard for swap or recharge.

A new demand signal for your platform

Delivery / fulfilment platforms: Smart Shelf gives you a predictive replenishment signal from the home — the order triggers before the customer thinks to reorder.

Incremental delivery orders from in-home signals
Higher attach rate on breakfast-fridge staples
A first-party demand signal — before the customer runs out.
Plug-in via REST API — you stay the delivery channel
We complement your channel, not compete with it

You're the store, the delivery channel, and the brand — we sharpen all three.

For vertical e-grocery

Forecasting

Demand signal on staples 48 hours before the household runs out.

Fulfilment

Optimise warehouse load & courier routes.

Waste

Fresh perishables delivered inside the window.

Private label

Real consumption data grows your own brands.

The run-out moment is the highest-intent touchpoint in the category.

For brands / CPG

Replenishment = media

Premium bottom-funnel moment: ran out → about to buy. Right of first offer.

First-party consumption data

Real consumption, not shipments: promo timing, triggers, true frequency.

Closed-loop measurement

Signal → triggered re-order → purchase. Proven incrementality.

Pilot Program

A focused, low-risk pilot that proves incremental basket, re-order conversion and wallet share on your own loyalty customers.

500

opt-in loyalty households

Q1–Q2 2027

deploy · run · measure

breakfast-fridge

anchor category

REST API

4–6 weeks to integrate

Measured incremental GMV vs. control group
First-party in-home consumption data
Pre-agreed success gates → contract
Co-funding options with breakfast-fridge CPGs

Pilot timeline

Integrate

Weeks 1–6

REST API + catalogue integration; align SKUs and success gates.

Recruit

Weeks 5–10

Invite 500 opt-in loyalty households in the anchor category.

Deploy

Weeks 9–14

Ship devices, onboard households and validate recognition.

Run & measure

Weeks 12–22

Track basket, re-order and retention vs. a hold-out control group.

Read-out

Weeks 22–24

Decision to scale, based on pre-agreed gates and commercial ROI.

Success gates

+10–15% incremental breakfast-fridge basket vs control
Re-order conversion ≥ target vs control
Retention ≥ control group
Recognition ≥ 90% on pilot SKUs

What we need from you

API / catalogue access
Loyalty household segment
Weekly ops point of contact

Team

Meet the team

Gregory Tublin

Founder & CEO

Retail strategy and partnerships.

Nikita Kaniukov

CTO Co-Founder / Hardware

Embedded systems, camera and sensor integration.

Mila Belova

US Partnerships Representative

US retail partnerships and pilot operations.

Aleksei Goriunov

ML/AI Engineer | Computer Vision Research & Development

Computer vision R&D and model deployment.

FAQ

Common questions from retail and innovation teams

You pay for the value you get — never the full £199. Three ways: £49 one-time, £5/month with nothing upfront, or with loyalty points and no cash out of pocket. Category partners and the retailer co-fund the rest through retail media.

Request the pilot

Tell us your category, region and integration stack. We’ll send a tailored pilot proposal.