How to get Gemini to recommend your brand

Gemini recommends brands Google can find and verify. Its answers are grounded in the Google Search index, crawled by the same Googlebot that feeds Search and rendered through headless Chromium. The single most damaging mistake brands make is blocking Google-Extended in robots.txt, believing it controls whether Google indexes their site. It does not. Google-Extended is not a crawler and has zero effect on Search inclusion. Blocking it only removes your content from Gemini's training and grounding data, which is the opposite of what you want.

Why Gemini skips your brand

Gemini's failure modes are its own:

1. Google-Extended is blocked by mistake. A privacy checklist somewhere told the team to block it, on the belief it stops Google from indexing the site. Google's own documentation says otherwise: Google-Extended is a policy-only robots.txt token with no HTTP user-agent, controlling only whether already-crawled content may be used for Gemini training and grounding. Blocking it does nothing for Search inclusion and quietly cuts you out of Gemini's answers.
2. JavaScript delay invisibility. Google renders JavaScript, but in two waves: initial HTML first, render later. Product content that only appears after JavaScript executes gets indexed late, and the delay is not guaranteed to resolve if render resources are blocked. This is different from Claude, where JS shells are hard-invisible; here it is a slow decay that quietly makes your data stale.
3. Thin local presence. For local questions, Gemini draws on public Google Maps information, and Google Business Profile feeds those Maps listings. A brand with an incomplete or outdated profile loses the queries where location matters. Google never says verbatim that Gemini reads your Business Profile, so treat this as the data chain it is: Maps information feeds local answers, and GBP feeds Maps.

Step 1: Unblock Google-Extended (unless you mean it)

Check [yourdomain.com]/robots.txt, replacing yourdomain.com with your actual domain. If nothing loads, create the file at the root of your site. Then look for a Google-Extended group. If you find a block like this and want Gemini to recommend you, change it to an allow:

User-agent: Google-Extended
Allow: /

Know what you are deciding. Allowing Google-Extended lets your content be used for Gemini training and for grounding (handing Google Search index content to the model at prompt time). Blocking it does not reduce your Search inclusion, ranking, or AI Overviews one bit. So the only question is whether you want your content usable by Gemini at all. If you keep the block, do it knowing it costs you recommendations, not knowing it protects you from indexing.

Do not look for Google-Extended in your server logs. It has no HTTP user-agent string and never requests anything, so any log line claiming to be Google-Extended is spoofed. Verify by fetching your own robots.txt and confirming which group is present.

Step 2: Confirm Googlebot can actually crawl

Google-Extended is the policy token, but Googlebot is the crawler that feeds Gemini's index. A block aimed at something else often catches it. First confirm robots.txt does not disallow Googlebot. Then check the edge, because the file is what you declare and the edge is what enforces: Cloudflare Bot Fight Mode and WAF bot rules can challenge Googlebot even when robots.txt allows it. In Cloudflare, open Security > Bots and confirm Googlebot is not being challenged. Then verify in your own logs and confirm the requests are genuine with reverse and forward DNS: the hostname must resolve to *.googlebot.com and forward-resolve back to the same IP. Finally, use Google Search Console: URL Inspection for a live test of a blocked page, Crawl Stats to see the trend, and request a recrawl after any robots.txt change.

Step 3: Beat the two-wave delay on product pages

View source on a product page, not the rendered page. If the price, description, and availability are in the HTML, you are fine. If they only appear after JavaScript executes, Google will get to them in the second wave: later, slower, and not guaranteed if render resources are blocked.

• Serve price, description, and availability as server-rendered HTML on product templates.
• Unblock render resources (JS, CSS, data endpoints) in robots.txt; a crawler that is allowed in but starved of resources still sees an empty page.
• Keep product URLs stable. Frequently changing URLs fragment your index presence.

You do not need to rebuild the site. Most platforms can serve pre-rendered HTML to crawlers without changing what shoppers see.

Step 4: Feed Maps for local queries

If you take local business, treat Google Business Profile as part of your AEO surface. Complete name, categories, hours, and ratings on the profile, because that data feeds the Maps listings that local answers draw on. Search your brand as a customer would: if the Maps panel is thin or stale, Gemini's local answers will be too.

What this guide does not cover

Being findable gets you into Gemini's answers. Staying there takes monitoring: Google re-indexes on its own schedule, your catalog drifts, and what Gemini says about you changes without warning. And answers are not orders; letting the agent complete the purchase is the transaction layer, which is separate work.

For the Claude version of this playbook, see our guide to getting Claude to recommend your brand.