Submit Feedback

The easiest feedback you'll ever send

AI agents are getting better every day, but they still get things wrong. A tool will return stale data, an answer won't match what you see in Seller Central, or your agent will call the wrong tool entirely. When that happens, you don't need to open a support form or write up a bug report yourself (though we do read every email that comes in to support@agentcentral.to).

The quickest way to send us feedback is to ask your agent to do it, right from the conversation where the problem happened.

It will pull together a structured report from the conversation: what you were trying to do, what happened, what should have happened, which tools were involved, and any error messages along the way. Before it sends anything, it shows you a summary and asks you to confirm. Nothing gets submitted without your explicit approval.

Before the report is stored, sensitive information is stripped out automatically. API keys, tokens, connector URLs, buyer names, addresses, phone numbers, emails, order IDs, tracking IDs, and full tool payloads are all redacted. ASINs in free text are aliased. If you left something sensitive in the conversation that you don't want included, you can add a privacy note explaining what was omitted.

The report comes from the conversation itself, which is what makes it different from a typical support form. Your agent has context you'd otherwise have to reconstruct from memory: which tools it called, what date ranges it used, what the responses looked like.

It assembles that into a structured packet with a timeline, evidence, and categorization (tool error, missing data, stale data, confusing tool behavior, permission issue, and a few others).

Every report gets read by a person on our team, and we use what you send to improve the product. If we need more context or want to let you know something was fixed, we'll follow up with you directly.