Feature
Recommendations that read your catalogue like a sales associate.
A recommendation carousel guesses from behaviour. An assistant can just ask — what is it for, who is it for, what is the budget — and then pick from your catalogue the way someone on the shop floor would.
What it does
Asks before it guesses
When a request is too vague to answer well, the assistant asks one clarifying question rather than returning ten products and hoping. One question usually turns a browse into a shortlist.
Explains the choice
Each recommendation comes with the reason it fits — the spec that matters, the price bracket, the use case. Shoppers convert on the reasoning, not the ranking.
Grounded in real products
Recommendations are drawn from your synced catalogue with live prices and stock. The assistant cannot invent a product, quote a price you do not charge, or suggest something discontinued.
Remembers the conversation
Budget, preferences and what has already been rejected carry through the conversation, so the third suggestion is better than the first instead of starting over.
Knows returning shoppers
The assistant recognises a returning shopper and can tailor suggestions to what they have browsed and bought before, within the privacy limits you set.
Adds to cart from the recommendation
"Add the second one" resolves to the right product and goes into your real basket, so the path from suggestion to purchase is one message long.
Questions about recommendations
How is this different from a "you may also like" block?
Can it recommend something that is out of stock?
Can I stop it recommending certain products?
Which platforms can it recommend from?
See what it recommends from your catalogue.
Ask it for a gift under £50, or the right product for a beginner, and see what it comes back with on your real products.