Start with consequential gaps

Prioritise missing facts that regularly block a buying decision: suitability, dimensions, compatibility, delivery, and returns. Use the recommendation detail to understand the suggested change rather than copying AI-written copy without checking it.

Read coverage before conclusions

Accessible, unavailable, sampled, and blocked pages mean different things. A scan based on a sample does not represent every product or competitor page. Old observations should be verified before acting on price or availability comparisons.

Close the loop

Make the approved change in the relevant shop page and use the available review or rescan workflow to check the result. Keep a record of what changed. An improved scan result is useful evidence, but it does not by itself establish a causal increase in sales.