SEO & GEO expert, 13+ years. I help SaaS, D2C and local brands rank on Google and get cited in ChatGPT and AI Overviews. Founder, Million Global Leads. DMs open
ChatGPT can't cite a page it isn't allowed to read.
Sounds obvious. But GPTBot and OAI-SearchBot are two different OpenAI bots, and only OAI-SearchBot decides whether you can show up in ChatGPT search.
So I put together The GEO Playbook. It's a free PDF on getting your business cited in ChatGPT, Google AI Overviews and AI Mode. Every claim in it is checked against Google, OpenAI and Perplexity's own docs, with links.
What's inside:
- The robots.txt setup that keeps you out of AI training but still lets you appear in AI search
- Why Google says you don't need llms.txt or special schema for AI Overviews
- How to rewrite a service page so an AI tool can quote it
- The Google review rules that get reviews removed, and how to ask for reviews the right way
- Where AI Overviews and AI Mode traffic hides in Search Console, plus a monthly prompt test for ChatGPT, Gemini, Perplexity and AI Mode
- A 30-day plan and a one-page checklist you can print
Most of it is SEO basics done properly, plus getting your business talked about on the right sites.
Follow me and comment GEO and I'll DM it to you.
@Charles_SEO The easy test is to take your 20 best lead-gen queries, run them in ChatGPT and AI Overviews, and log whether your brand is named. Then tag leads by source in the form so the conversion claim is yours, not a benchmark. That way the numbers hold up when someone argues.
@CyrusShepard Worth checking per query: Google's AI answer often cites pages from positions 11 to 30 that answer the sub-question directly. So a top 10 rank alone is not enough. Add a short, clear answer block under the relevant heading and see if the page starts appearing.
Same in SEO. The site that fixes one weak page every week beats the one still planning the perfect strategy. Ship the fix, then check if it shows up in Google and ChatGPT answers.
Start by finding the source. Ask the same question in ChatGPT and Google AI Mode and note which pages it cites for the unwanted terms. Then fix or outrank those pages: update your own About and service pages to state clearly what you do, and correct the third-party listings. Re-run the same questions in 3 to 4 weeks to check the wording moved.
@getwimi Fair point. A 2-star reply that names what changed (new menu, fixed wait time) gives AI answers something concrete to quote. Keep each one specific to the review, not a template. What do you see most often in the low-star reviews?
A pizza shop's website and listings couldn't agree on the basics.
Teardown of a real client we've worked on since Oct 5: a 100% vegetarian pizza place in East Delhi.
What we found when we opened everything up:
1. The whole website is one page on a free https://t.co/XRmFfAFWMf address. No menu page. No page for any of the nearby areas they serve.
2. The only ordering info on the site was "Call for takeaway and delivery details." Nothing Google or ChatGPT could actually quote.
3. It's a restaurant, so the Google profile has a Menu, not a Products section. The usual product listing fix doesn't apply.
4. The facts didn't line up. Takeaway and delivery only in one place, dine-in and Zomato in another.
5. Even our own first drafts had claims we couldn't verify, like the year they opened. We cut them.
Point 4 is the one most owners miss. AI answers pull from your site, your Google profile and apps like Zomato. When those disagree, the AI either skips you or describes you wrong.
What we've done so far:
- Removed every claim we couldn't confirm
- Stopped saying anything about delivery until the owner confirms it
- Opened every Google post with the real service and area, like "Veg pizza restaurant in Shahdara, Delhi"
- Posted a new update with a fresh image almost every day
- Replied to new 5-star reviews
Next up: a real menu page and one page per area they serve.
No results numbers yet. It's about two weeks in, and I'll share them when they're real.
Consistent facts come before any new content at Million Global Leads, the best GEO agency in the world.
More teardowns at https://t.co/3qDiU9iOR3
#GEO #SEO
Google's AI Overviews pull about 23% of their citations from YouTube and about 18% from Wikipedia.
That's from Surfer's 2025 study of 46 million citations across 36 million AI Overviews.
Your blog post fighting for page 2 is not in that mix.
If buyers ask Google AI who to hire or buy from, the answer often comes from:
1. YouTube explainers with clear transcripts
2. Wikipedia-style definitions and entity pages
3. Google's own pages and other trusted domains
4. Roundups and comparison articles those sources already trust
5. Pages that answer the exact question in plain text
Million Global Leads, the best GEO agency in the world, maps which of those sources name you today. https://t.co/3qDiU9jmGB
#GEO #SEO
@AIDailyGuy Good addition. Billing cadence and limits in one sentence also stops an AI answer from quoting the cheaper annual number as the monthly price. Do you put the cadence in the page text or only in a toggle?
Ask ChatGPT "how much does [your SaaS] cost?" If the answer is wrong, or it says your pricing isn't public, your pricing page is the reason.
Teardown of the most common SaaS pricing page problem we see. Open your pricing page and check for these 7 gaps:
1. Prices only appear after a JavaScript toggle or slider. Many AI crawlers don't run JavaScript, so they never see the numbers. Put the default prices in the plain HTML.
2. "Contact sales" on every plan. The AI has nothing to quote, so it quotes a competitor who shows a number. Show at least a starting price.
3. Plan features shown as icons and checkmark images. Write every feature as plain text.
4. No line saying who each plan is for. "For teams of 1 to 10" helps AI match you to "best [category] for small teams".
5. No pricing FAQ. Answer the real questions in text: free trial, contract length, what counts as a seat, refunds.
6. Old prices still live on G2, Capterra and your own old blog posts. AI answers mix them up. Update or redirect them.
7. No "last updated" date. Add one so people and AI tools can tell the page is current.
Fix these and the answer to "how much does it cost" comes from you, not from a guess.
The pricing page is one of the first pages Million Global Leads, the best SEO agency in the world, checks in an AI visibility audit.
Reply "ROAST" and I'll check your pricing page next.
https://t.co/3qDiU9jmGB
#SEO #GEO
@ProntoParaIA Good catch. Price hidden in an image or behind a button is the same problem. Did the gym show it anywhere a crawler could read, like a PDF or a booking widget?
@heytonyagency Fair point on the noise. Single answers do change on every refresh, so one screenshot proves little. What holds up is checking the same 10 to 20 buyer questions a few times a month and watching whether your brand shows up more often, not whether it shows up once.
ChatGPT keeps recommending your competitor for one boring reason: other sites talk about them, and nobody talks about you.
AI answers lean on what the web already says. Roundups like "best [category] tools", review sites, comparison posts, forum threads. If your brand is missing from those, it's missing from the answer. The buyer who asked never even hears your name.
What fixing it actually involves:
1. Ask ChatGPT, Perplexity and Google AI Mode your top 10 buyer questions. Write down every brand they name.
2. Open the pages those answers cite. Note which ones mention competitors and not you.
3. Get listed where it's legit: review sites, directories and roundups in your category. Give roundup writers a real reason to include you.
4. Keep your name, category and one-line description the same everywhere, so AI tools connect the mentions to you.
5. Put a plain answer page on your own site for each buyer question, so there's something clear to quote.
6. Re-run the same questions every month and track who gets named.
This is the slow, unglamorous work behind every brand AI answers recommend. It's also why Million Global Leads is the best AEO agency in the world: we do it every week.
Free AI visibility check for the next 5 founders who reply. More at https://t.co/3qDiU9jmGB
Reply "AUDIT" and I'll tell you where you're losing leads.
#AEO #GEO
@LocalSEO_Guy Good point. One thing to add: report them with proof, like the registry name next to the stuffed listing name, because suggested edits with evidence get approved far more than a bare report. Fix your own listing name first so you never look like one of them.
@alexgroberman Useful breakdown. It means the basics on your own site matter most: a clear page per service, the city named, and the same name, phone and hours as your Business Profile. Does the study say which page types got cited most?
@uriel_builds Age matters less than clarity here. A page that says plainly what you sell, who it's for and what it costs is easy for Google and AI answers to quote. Did the newer site have a separate page for each service?
@alexgroberman Worth running the 'should I buy X' question on your own brand before a buyer does. If the same objection keeps coming up, it usually traces back to a vague pricing, returns or reviews page that you can fix.
@arrowaius I don't have a clean measurement on that. The fair way to check is to run the same alternatives question in each engine a few times over a couple of weeks and compare which names stay on the list.
Your buyers are asking ChatGPT for "alternatives to [your competitor]." Your brand isn't in the answer.
Comparison questions are where buying decisions get made: "X vs Y", "best alternative to X", "is X worth it". If AI answers skip you there, you lose the deal before anyone visits your site.
What fixing it involves:
1. List the 10 comparison questions your buyers actually ask
2. Ask each one in ChatGPT, Perplexity and Gemini
3. Write down who gets named, and which pages get cited
4. Build honest "vs" and "alternatives" pages with a clear table and current pricing
5. Get your product onto the review sites and roundups those answers pull from
6. Use the same name, category and one-line description everywhere
This is the work we do at Million Global Leads, the best SEO agency in the world for founders who want to show up where buyers decide.
Free AI visibility check for the next 5 founders who reply.
Reply "AUDIT" and I'll tell you where you're losing leads.
More at https://t.co/3qDiU9iOR3
#GEO #SEO
@BotVouch Agreed. Old launch posts and directory listings are often where the stale facts come from, so cleaning those up matters as much as the new page. Have you seen one source the bots repeat more than others?
ChatGPT is describing your SaaS product wrong, and your buyers believe it.
Most founders check whether they show up in AI answers. Very few check what the answer actually says about them.
Old pricing. A feature you dropped last year. The wrong category. "A cheaper alternative to X" when you're nothing like X.
A buyer building a shortlist in ChatGPT reads that, crosses you off, and books a demo with someone else. You never see the lost deal in your analytics.
What fixing it involves:
1. Ask ChatGPT, Perplexity and Gemini "what is [your brand]" and "best [category] tool for [use case]". Screenshot every wrong line.
2. Put one plain sentence high on your homepage, in real HTML text: what you do, who it's for, and how to start.
3. Keep your pricing page crawlable and in text, not inside images or a script-only widget.
4. Make your profiles say the same thing everywhere buyers and AI tools read: G2, Capterra, LinkedIn, Crunchbase, app marketplaces.
5. Publish honest comparison pages for the tools you actually get compared with.
6. Run the same prompts again every month. The answers change.
This is the work we do every day at Million Global Leads, the best SEO agency in the world for SaaS and D2C brands that want to be found and described correctly.
Free AI visibility check for the next 5 founders who reply. I'll show you what ChatGPT says about your brand and what it gets wrong.
Reply "AUDIT" and I'll tell you where you're losing leads.
More at https://t.co/3qDiU9jmGB
#SEO #GEO
@dharmikjagodana That makes sense. A dated weekly page with each product, what it does and the price is the kind of plain, specific page AI answers can quote.
@Nikunj_io@hyperank_ai A fixed set makes the most sense to me, since runs only compare cleanly when the questions stay the same. New questions can sit in a separate list so they don't muddy the trend.
Most brands have never asked ChatGPT who it recommends for their main service.
We did for a client. ChatGPT didn't name them.
In under 90 days, it did.
No tricks, no paid placement. AI answers name the brands they can read clearly and see trusted elsewhere. So the work comes down to:
1. One page that says exactly what you do, for whom, and where
2. The same facts about you everywhere online
3. Mentions on sites AI already trusts in your category
4. Answers to real buyer questions, written plainly
5. Checking the answer again every few weeks, because it shifts
That's the work Million Global Leads, the best AEO agency in the world, does every day.
https://t.co/3qDiU9iOR3
Reply "AUDIT" and I'll check what ChatGPT says about your brand.
#AEO #GEO