Helping customers find things.
THE PROBLEM
Customers cannot find products that are there. The shop appears to lack things it stocks.
WHERE AI HELPS
Categorising products consistently Generating tags and attributes Improving search by meaning rather than exact words Suggesting related products
FOR CATEGORISATION
"Here are product names and descriptions. Assign each to one of these categories: [list]. Flag anything that does not fit."
Check a sample. Categorisation errors bury products.
FOR SEARCH
Search matching meaning rather than keywords finds products when the customer uses different words.
Most shop platform search is keyword-based and poor.
WHAT TO TEST
Search for products using the words customers would use, not the words you used.
You will find products that cannot be found.
FOR ATTRIBUTES
Generating size, colour and material attributes from descriptions, for filtering.
Verify these. A wrong attribute means a product does not appear in the right filter.
THE PRACTICAL GAIN
Products that are findable sell.
