How AI stocktaking works
Open StockLens on any smartphone, tablet or laptop with a camera. Point it at a shelf, fridge or back-bar. The AI detects each bottle, matches it to your product library and reads the fill level by analysing the liquid line. You confirm or adjust on screen and the count is saved.
There is no specialist hardware. There are no barcodes to scan. The same workflow handles spirits, wines, beers, soft drinks and post-mix.
Why operators move from manual stocktakes to AI
Manual stocktakes are slow, expensive and error-prone. A typical 200-line bar count takes one person three to four hours, twice a month. Mistakes feed straight into variance reports and skew GP calculations.
- Cut stocktake time from hours to minutes
- Reduce human counting errors
- Free up managers for service, not spreadsheets
- Standardise counts across multiple sites
- Get a defensible audit trail with photos attached
What StockLens AI counts
Our model is trained on hospitality SKUs across major UK and European spirit brands, wine, beer, cider, RTDs and mixers. It also handles glassware levels, kegs and consumables.
Accuracy and audit trail
Every AI count is stored alongside the original photo. If a variance is queried later, you can pull up the exact image used for the count and re-verify. This is impossible with paper stocktakes.
Built for hospitality, not generic retail
StockLens AI was built specifically for bars, pubs and restaurants. Sections, cellar transfers, AI-captured sales reports and GP reporting are all designed around the way hospitality actually trades — not adapted from a warehouse system.
Hospitality use cases
- Independent gastropubs reducing weekend manager hours
- Multi-site bar groups standardising counts across venues
- Hotel F&B teams running fortnightly stocktakes across multiple outlets
- Restaurant chains tracking beverage variance against EPOS sales