On 17 November 2026 Google published a help page called “Allow Google-Agent traffic”. Google-Agent is the name for agents that run on Google’s own infrastructure, browse websites and take actions when a user asks them to. If your website sits behind a CDN, a firewall plugin or a bot-detection service, those tools may treat such a visitor like a scraper and block it. Google’s page explains how to recognise the traffic and let it through while keeping your other protections. For a small business with a WordPress site, this is a one-hour check worth doing, because a blocked visitor is a lost enquiry you will never see in your reports.
What Google-Agent is, in plain words
Googlebot reads pages so they can appear in Search. Google-Agent is different: it acts for a person. According to Google’s documentation, these agents browse the web and perform actions on a user’s request. Search Engine Roundtable notes that the agent itself is not new, since it was covered in March, but the technical guide for publishers is.
Think of a customer who asks an assistant to find a dentist with a Saturday slot, compare prices, or fill in an enquiry form. The assistant may open your website, move through pages and submit something. If your security layer shows it a CAPTCHA or an error page, the assistant moves on to your competitor. Google does not promise how often this happens, and we are not claiming it will happen to your site. The point is that the door should not be locked by accident.
How Google says you can recognise it
The help page lists three ways to identify the traffic:
- The
User-Agentheader contains “Google-Agent”. - The source IP address appears in Google’s published range file for user-triggered agents.
- Google is testing Web Bot Auth with the identity
https://agent.bot.goog. Participating requests are signed, and the public keys are published at a well-known directory on that domain.
The page’s advice is layered. If your CDN or bot-protection provider can verify Google-Agent natively, use that feature. If not, check the user-agent text, confirm the IP is in Google’s range file, and verify the signature when one is present. Once a request is verified, skip bot challenges for it. Never allow traffic on the user-agent text alone, because anyone can fake that header.
Six ways sites still block it after “allowing” it
This is the most useful part of Google’s page. It lists troubleshooting points that we see often when auditing small-business sites:
- Rule order. Your allow rule must run before broad rules that block by network, managed bots or AI agents. An allow rule at the bottom of the list never fires.
- Challenges. Exempt verified requests from CAPTCHAs and JavaScript challenges, including the small CAPTCHAs that appear when someone searches or submits a form.
- Multi-step journeys. The allowance should cover search, filters, detail pages, cart and booking actions, and the backing API endpoints, such as XHR, GraphQL or
api.subdomains. Allowing only your home page is not enough. - Rate limits. Exempt verified ranges from per-IP limits, because many users can share the same outgoing addresses.
- Range updates. If you copy IP ranges into your rules, keep them in sync with Google’s latest file.
- Proxies and load balancers. Make sure they pass on the client IP, the user-agent and the signature headers instead of stripping them.
Website Development · FSK Digital Marketing
A fast website built to bring in calls.
From ₹24,999 one-time
A one-hour check for your WordPress site
You do not need to be a developer to start. Work through this in order and write down what you find.
- List your protection layers. Do you use Cloudflare or another CDN? A hosting firewall? A WordPress security plugin? A CAPTCHA on forms? Many Indian small-business sites have two or three layers and the owner knows only one of them.
- Find the bot settings. Look for “bot fight mode”, “AI bots”, “managed challenge” or similar names. Note which are on.
- Read your provider’s help for Google-Agent. Google says some providers verify it automatically, so check before writing custom rules. If the provider’s documentation does not mention it, ask their support.
- Check your security event log. A block at the CDN may never reach your own server logs, so a clean WordPress log does not prove nothing was blocked.
- Review your forms. If your enquiry or booking form uses a CAPTCHA, think about what a verified assistant would see.
- Do not loosen everything. The aim is a precise exception for verified Google traffic, not turning off protection. The Google page itself stresses keeping other security controls in place.
Security is a real concern right now. The FBI recently linked compromised WordPress sites to the Gootloader malware, which we covered in our WordPress security checklist. So any change should be made carefully, ideally on a staging copy.
Why a blocked visitor is invisible to you
This is the awkward part. When a firewall turns someone away, your analytics usually record nothing, because the page never loaded. Your WordPress dashboard looks healthy and your enquiry inbox is simply quieter than it should be. That is why Google’s advice to read the CDN or firewall event log matters more than reading your site logs.
A simple habit helps: once a month, open the security events in your CDN or hosting panel and scan the blocked requests. Look for patterns such as a rule blocking an entire country, a challenge on every form, or an API path that returns errors. Fix the rule that is too broad, not the symptom. If a rule was added during a past attack and nobody remembers why, test it on a copy of the site before removing it. Keep a short written list of every rule you change, with the date, so you can undo it.
What FSK does and does not promise
Allowing Google-Agent will not raise your rankings, and nobody can honestly promise that. It simply removes one possible reason a legitimate visitor fails to complete a task on your site. If your site is slow, outdated or built on stacked plugins, it is often easier to fix the foundation than to patch rule after rule. Our WordPress website development service starts at ₹24,999 and covers clean builds with sensible security and fast pages. If you only need your Google presence checked, the Google Business Profile audit at ₹1,499 is a smaller first step.
Frequently asked questions
Is Google-Agent the same as Googlebot?
No. Googlebot crawls pages for Search. Google-Agent is the user agent for agents on Google’s infrastructure that browse and act when a user asks. They are documented separately.
Can I identify it by its user-agent text alone?
Google’s page lists the user-agent as one signal, alongside the published IP range file and, where present, a signature under Web Bot Auth. Because headers can be faked, verify more than the text before skipping any challenge.
Will this help my business appear in Google Maps or Search?
Google does not say so, and neither do we. This is about not blocking legitimate traffic. Your visibility still depends on the usual work: an accurate profile, helpful pages and real reviews.
If you are unsure what your CDN, plugin and hosting firewall are doing, start with the one-hour check above and note the questions. When you want an expert to go through it, get in touch with FSK.
Website Development · FSK Digital Marketing
A fast website built to bring in calls.
From ₹24,999 one-time

