SARTOAI INTELLIGENCE / ENGLISH ARCHIVE
How do customers find products on your site?
Distinguish a genuinely missing product from a product customers cannot find. Start with product records, availability and existing search settings before testing AI-assisted intent matching.
2026-10-13 · English edition: 31 October 2026
Download full English PDF ↗ · Original Turkish PDF ↗
In stock.
Missing from search.
The customer searches for “bol paça”; the catalogue says “wide leg”. Is a missing product really the problem?
Before buying more products or increasing advertising spend: a search review card that identifies products customers cannot find.
The business problem / the technology / a tool you can use
The same trousers.
Two ways to describe them.
Illustrative example. Not a test of a real store.
If this search returns no results, first investigate search behaviour or the product record, rather than assuming that the product is missing.
- WHAT THE CUSTOMER SEARCHES FOR
- “Black wide-legged trousers” Original query: “Siyah bol paça pantolon”
- HOW THE CATALOGUE DESCRIBES IT
- “Wide leg trousers” Colour: black / available for sale
Before AI,
find out what is missing.
SartoAI’s proposed sequence. Not every search problem needs AI.
AI may help connect “bol paça” with the relevant product. Do not count navy as a correct result when the customer asks for black.
- THE PRODUCT IS NOT VISIBLE
- Check publishing and inventory settings.
- THE WORDING DIFFERS
- Correct product information and search settings.
- THE INTENT IS NOT UNDERSTOOD
- Test AI-assisted meaning matching.
No-results search
review card
E-commerce + product team. A completed illustrative example.
Complete the card using your own queries. Before saying “in stock”, check the relevant sales channel and the requested size.
- SEARCH QUERY
- Black wide-legged trousers
- EXPECTED PRODUCT
- P-104 / wide leg / black
- OBSERVED DURING REVIEW
- Available for sale, but missing from search.
- FIRST CHECK
- Product name + search matching
A first project with SartoAI:
which search, which correction?
Proposed project scope. Not a ready-made product or demo.
If your existing software is sufficient, adjust its settings. This review establishes whether custom development or integration is needed.
- INPUT
- 20 searches with no results + the current product list
- REVIEW TOGETHER
- Is a suitable product actually available? Why is it not being found?
- YOUR OUTPUT
- A correction list and test examples for each search
Test the search
you already have.
MARIA SARTO / RECOMMENDATION CARD
First conversation output: a list showing what needs to be corrected for each search.
The right product, not simply more results.
- OUR RECOMMENDATION
- Test your existing search first.
- WHAT IT IS BASED ON
- Shopify’s search guides [S1-S2]. The card and proposed sequence are SartoAI recommendations.
- YOUR FIRST STEP THIS WEEK
- Select 20 searches that returned no results. Ask the product team to use the card on page 4 to check whether a matching product is in stock.
Sources and archive context
These guides describe search behaviour and options for AI-assisted meaning matching, not a new launch or measured sales results. Features depend on the plan, theme and search provider. Test Turkish performance against your own queries. The trouser example, review card and proposed sequence are SartoAI designs.
Maria Sarto · SartoAI Intelligence. Sources retain their original publication context; these articles do not claim measured customer results.