Are you verifying that products are readable by AI shopping

With more shoppers starting product discovery in AI assistants (ChatGPT, Copilot, Gemini, Perplexity), I’m curious how teams are handling the pre-visibility problem.

Specifically:

How do you check whether product data is actually interpretable by AI systems (not just indexed)?

Are you relying on schema validation, feed checks, manual prompting, or something else?

Have you seen cases where visibility tracking showed “nothing,” but the root cause was unreadable or ambiguous product data?

Interested in practical approaches, failures, and lessons learned