Sanikey
Foundations first. Then, AI.
B2B distribution
Project information
Sanikey has been supplying medical and hygiene products for twenty years. It sells on a flat-rate basis to care homes, hotels and hospitals: each centre has its own monthly balance, catalogue and prices.
All of that lived in an old PrestaShop, held together with workarounds. Every pricing rule was a patch. And the monthly balance was a headache.
They asked us for machine learning. We told them: not yet.
AI learns from data, and there was no reliable data. And on a base that was falling apart, anything clever ends up as just another workaround.
So foundations first. A private store that adapts itself to each centre, with its own catalogue, prices and balance. A portal for each client's supervisors. And a panel so Sanikey can run it all without touching code.
On top of that, what did make sense from day one: the suggested order. After three orders, the store proposes the next one in full, with the mandatory products and fitted to the balance. For a centre ordering lots of items every month, that's time saved and mistakes that don't get made.
Today 137 centres use it and it has handled 61,540 orders. With that data, now it's time: AI is already in. It fine-tunes the suggested order and forecasts what each centre will use. And we keep shipping improvements.
- Services
UX/UI
B2B online store
Custom software
AI and automation
Ongoing development
Project summary
They wanted machine learning on top of a PrestaShop full of workarounds. We laid the foundations first: a store that applies each centre's rules by itself and suggests its order for the month. Today, 137 centres, over 61,000 orders and, now, AI.
Impact achieved
- 137
- centres
- 61.540
- orders
Predictive ordering
After three or more orders, the store suggests the next order in full. It looks at how often the centre buys each product and in what quantity, and compares it with the pattern of the most efficient buyers at the same client. It also takes into account the history from before the website, imported from the old system.
Then it adds the mandatory products and adjusts the order to fit the month's available balance. The buyer reviews it, changes whatever they like and moves it to the basket in one click.
And now, with two years of orders behind it, AI fine-tunes the suggestion and forecasts each centre's consumption.
Cascading prices
Every centre buys on its own terms, and the store applies them automatically at login. Each product's final price goes through four adjustments, one after another: buyer, client, segment and service.
The centre sees its Sanikey price next to the retail price, and the Sanikey team changes the terms from the admin panel, without touching code for each new client.
A catalogue for each centre
Sanikey sells to care homes, hotels and hospitals, and each has its own terms: which products it can see, at what price and how much it can spend a month. Before, all of that lived in a PrestaShop full of patches.
Now the store adapts itself to whoever logs in. Products, categories and services are hidden or allowed by centre, client or segment, and everyone sees only what's theirs, with their balance for the month. Adding a new client takes no code: it's set up from the panel.
“We wanted artificial intelligence and they told us not yet: first, a store that worked. They were right. Today 137 centres use it and the monthly order almost places itself.”
Next project
RTTH
From knowledge to impact