Case studies
Four clients. Work running every day.
A supermarket chain, Aayra Crown, Vital Drugs and Thebo. Every piece below is labelled: live means a client uses it every day.
One supermarket chain, three AI tools.
A premium supermarket chain, Bangkok. Shelf checks that used to take a day now run overnight. The three screens a chain sees, rebuilt with sample data.
Photo · 21:48 · staff phone| Store | What's wrong | ฿/day | Status |
|---|---|---|---|
| Store 04Aisle 7 | Pringles, 2 facings empty | ฿2,400 | To fix |
| Store 11Dairy B3 | Greek yoghurt 500g gone | ฿1,900 | Fixed |
| Store 04Aisle 7 | Coke can in the Pringles slot | ฿380 | To fix |
| Store 17Coffee B2 | Beans 1kg out of stock | ฿860 | Checking |
| Store 11Wine C2 | Price tag missing | ฿150 | Fixed |
Sales
฿284,691
+12% on last quarter
Gaps fixed
143
across 18 stores
Facings matching
94%
up from 81%
Weekly sales, indexed
Shelf checks started week 9
By weekday
Shelf intelligence
Staff photograph the shelves on a phone. By morning, each store has a worklist of gaps, ranked by the sales they are costing.
LiveThai-language shopping concierge
Answers shoppers in Thai, suggests pairings from the real catalogue, and hands over to staff when it should.
LivePromo banner studio
Builds promotion banners from the live campaign list and real product photos.
LiveBefore
Someone walked each store with a printed planogram and a clipboard.
Now
Staff photograph the shelves on a phone. Each photo is matched against the catalogue and that shelf's sales, and every store opens with a fix list ranked by the money at stake.
32,141
products it recognises
Buyers walk a 224-home project before it's built.
Aayra Crown: three towers and 224 homes in West Bengal. One tool is live, one is being built, one is a demo, and all three started from the same architect's drawings.
Step 1
The architect's drawings
Three towers, 224 homes, exactly as approved.
Step 2
Pick a tower and a floor
On the real plot, with the real sun path.
Step 3
Pick a flat
Size, facing and availability on one card.
Step 4
Walk inside
Furnished rooms, from a link on the phone.
3D buyer walkthrough
Buyers choose a tower, a floor and a flat, see the real sun at their window and walk through furnished rooms, from a link on their phone.
LiveWhatsApp enquiry reply
Answers flat enquiries from the live inventory and books site visits. Being built for the same kind of launch.
In buildQuantity takeoff
Floor plans measured into a bill of quantities, tested on the same project's drawings. Any number money depends on needs an engineer's sign-off.
DemoBefore
Buyers chose a flat from floor-plan PDFs and brochure renders.
Now
One link opens the whole project in 3D, built from the approved drawings: pick a tower, a floor and a flat, then see the real sun at the window.
224
homes, three towers, one link
Vital Drugs: an ERP turned into a reorder list and a margin view.
Vital Drugs, a pharmacy wholesaler in Dubai. Owners, managers, the buying team and field reps work from the same numbers, built on the ERP the company already ran.
BI dashboard
Sales, margin, procurement and inventory on one screen, built on the ERP's own exports, for owners, managers and sales reps.
LiveReorder console
The buying team's daily worklist: what to order, from whom and how much, with expiry risk and every order tracked until it is received.
LiveBefore
Owners waited on spreadsheets for margin. Buyers reordered from memory and a printed stock list.
Now
Staff drop in the standard ERP exports. Every morning the buyers get a ranked reorder list with the cheapest proven supplier, and the owners see margin, expiry risk and stock-outs on one screen.
4 roles
owners, managers, buyers and reps on one set of numbers
Supplier price lists into product records, without the retyping.
Thebo, a Dutch catering and refrigeration equipment distributor selling through six webshops. Nothing goes live until a person approves it.
Listino prezzi 2026 · pag. 14
Armadio refrigerato GN 2/1
Capacità 650 L · Temp. -2/+8 °C
Dimensioni 740×830×2010 mm
Prezzo listino € 2.480,00
Product record · waiting for approval
approved by a person
Before
The team retyped long supplier PDFs, often in Italian, into each product record.
Now
It copies every detail into the product record, shows where on the page it found it, and waits for a person to approve it.
99%
less data-entry time
Three more industries, playbooks ready.
The same kinds of tools, mapped to the work in each. Every playbook names the job, the systems, the target and how we grade it.
Tell us the job your team hates. We'll tell you if AI can do it.
A 30-minute call with the people who would build it. No technical knowledge needed: just describe the job, how often it happens, and how you check it's done right.