Skip to content
regensbot
DEV-AIF-04Should4 of 6 in AIF

Do not block the AI agents you want as visitors

Level
Should

Recommended: skip it only for a reason you can name.

Documentation

Said at Search Central Live

Said or shown at the event and not in Google's documentation: read it as said at Search Central Live, not as documented policy.

Covered

Assisted

Proposed with the evidence. A person confirms it.

Why it matters

Said at Search Central Live: Google's duplication talk closed with the advice not to block agents, and AI agents browsing for users run into the same bot walls that sites build against scrapers and may go elsewhere, for example to buy the product from another shop.

How to meet it

Decide per agent which ones to allow, tune bot protection to challenge abusive patterns rather than every automated browser, and if a challenge is needed send it with 503 (DEV-SRV-02).

How to test it

WAF and bot-management logs show no blocks of agents the team decided to allow.

How Regensbot covers it

AssistedNeedsYour server logs

Regensbot proposes a verdict with the evidence; a person on your team confirms it.

The evidence comes from your server logs once you connect them.

regensbot logs push keeps only lines from Google's and published agents' IP ranges, so no visitor IP leaves your systems. Googlebot is verified by IP and DNS.

How verdicts workWhat you connect

Sources

Google documents00

No Google document is cited. This requirement was said at Search Central Live and is not documented policy.

Claims in the knowledge base03

  1. D2-C397UndocumentedHandling web duplicationShown on screen
  2. D2-C377UndocumentedHandling web duplicationSaid on stage
  3. D2-C378n/aHandling web duplicationThe author's interpretation

Claim IDs refer to the Search Central Live 2026 knowledge base. The knowledge base (opens in a new tab)

Undocumented
Said or shown at the event, not in the documentation.
n/a
Nothing to verify: a quotation, framing, opinion or analysis.

Audit

Check it on your site.

One audit checks a URL against all 138 requirements and returns one finding for each, with its source. One URL decides 66; for each of the others, the finding says which source to connect.