Ranking in ChatGPT is not quite the right phrase. There is no position one. When someone asks a question, the AI writes an answer and names a few businesses or sources. Your goal is to be one of them, consistently, for the questions your customers ask.
This is the practical playbook. The generative engine optimisation guide explains the theory behind why AI engines cite what they cite. This post is the checklist.
Key takeaways
- ChatGPT answers from two places: what it learned in training and what it searches live. You can influence both.
- Check your robots.txt allows AI crawlers. Many sites block them by accident.
- Build a list of target prompts the way you would build a keyword list.
- Third party mentions drive AI recommendations more than your own site does.
- Track AI referral traffic in GA4 from day one so you can prove return on investment.
The short answer
To get recommended by ChatGPT and cited in Google AI Overviews: make sure AI crawlers can access your site, get indexed in both Google and Bing, structure your content so key answers can be lifted cleanly, back your claims with specific data, and earn consistent mentions of your brand across reviews, directories, press and forums. Then track target prompts monthly and keep closing the gaps.
How ChatGPT decides what to mention
ChatGPT draws on two sources of knowledge.
1. Training data. What the model learned from the web before its cutoff. If your brand was widely and consistently discussed, the model knows you. This is slow to influence and reflects your reputation over years.
2. Live search. For many queries, especially recent, local or commercial ones, ChatGPT searches the web in real time, reads results and cites them. This is where you can make a difference quickly.
OpenAI runs separate crawlers for different jobs. Its crawler documentation lists them: GPTBot collects data that may be used for training, OAI-SearchBot finds pages to show in ChatGPT search results, and ChatGPT-User visits pages when a user's request needs it. Blocking OAI-SearchBot means you cannot appear as a cited source in ChatGPT search. Plenty of businesses have done exactly that without knowing.
How Google AI Overviews decide what to cite
AI Overviews draw on Google's own index. Google has said there is no special markup or trick required: pages need to be indexed and eligible to appear with a snippet in normal search. Beyond that, Google runs several related searches behind the scenes and pulls the passages that best answer each part of the question.
That is why strong SEO is the entry ticket but not the whole game. Your page has to be in the index, and then your passage has to be the clearest, most trustworthy answer to one part of the question.
The 10 step playbook
Step 1: Check AI crawler access
Open your own robots.txt and look for blocks on AI crawlers. Security plugins and CDN settings sometimes add them by default. If you want to appear in AI search, it should allow them:
User-agent: OAI-SearchBot
Allow: /
User-agent: ChatGPT-User
Allow: /
User-agent: PerplexityBot
Allow: /
User-agent: ClaudeBot
Allow: /
User-agent: Google-Extended
Allow: /Whether to allow GPTBot, which is the training crawler, is a business decision. For most small businesses being part of the model's general knowledge is a benefit, so I usually recommend allowing it. Also check your CDN for bot blocking settings that override robots.txt.
Step 2: Get indexed in Bing
The most overlooked step in LLM SEO. Several AI tools draw on Bing's index, and plenty of small business sites have never been verified there.
- Verify your site in Bing Webmaster Tools. You can import straight from Search Console in a couple of minutes.
- Submit your XML sitemap.
- Enable IndexNow so Bing hears about new and updated pages immediately.
Step 3: Build your target prompt list
Prompts are the new keywords. List 20 to 50 questions your ideal customers would ask an AI, grouped by intent:
| Prompt type | Example, for a Birmingham accountancy firm |
|---|---|
| Recommendation | Who are the best accountants for small businesses in Birmingham? |
| Comparison | Should I use a local accountant or an online accounting service? |
| Problem | How do I reduce my tax bill as a limited company director? |
| Cost | How much does an accountant cost for a sole trader in the UK? |
| Brand | Is this firm any good? |
Recommendation and brand prompts are where revenue sits. Problem and cost prompts are where you build authority and get cited.
Step 4: Run a baseline audit
Ask each prompt in ChatGPT, Gemini, Perplexity and Google, to catch AI Overviews. Use a logged out or fresh session so personalisation does not skew the result. Record whether you are mentioned, whether you are cited with a link, which competitors are named, and which sources are cited.
The sources cited column is the valuable one. It tells you exactly which websites the AI trusts for your topic. If it keeps citing a particular directory, a forum thread or a trade publication, that is where you need to be.
AI answers vary between runs, so ask each prompt a few times and record how often you appear rather than a single yes or no.
Step 5: Create content that answers your prompts directly
For each problem and cost prompt, make sure you have a page that answers it better than anything currently cited:
- A direct answer in the first two sentences
- Specific numbers, ranges and timeframes
- A clear date showing it is current
- Tables for comparisons and pricing
- FAQ sections using the exact phrasing of your prompts
Step 6: Build honest comparison content
AI engines lean on comparison content because recommendation questions are comparisons in disguise. Pages weighing one option against another get cited heavily.
Be genuinely fair. A comparison that says you are best at everything reads as marketing and gets passed over. Saying who you are not right for is what makes the rest of the page credible.
Step 7: Earn third party mentions
This is the step that separates brands AI recommends from brands AI ignores. When an assistant answers a best X in Y question, it leans on what other sites say, not what you say about yourself. Prioritise based on what your baseline audit showed was being cited:
- Reviews on Google and the platforms relevant to your industry
- Directory and best of listings the engines are already citing for your prompts
- Industry publications and local press, which is digital PR work
- Forums, through genuine helpful participation, never disguised promotion
- Podcasts, webinars and guest articles where your expertise gets discussed
For local businesses this overlaps heavily with local search. The local SEO guide covers the citation and review side in detail.
Step 8: Make your brand information consistent
If your website says you serve five counties, your Google profile says three, and a directory lists an old address, the AI has conflicting information and is less confident recommending you.
- Audit your name, address, phone, services and service areas across every listing
- Add Organization schema with sameAs links to your official profiles
- Make your About page state plainly who you are, what you do, where and since when
Step 9: Keep content fresh
Engines that search live prefer current information, especially for anything involving prices, rules or best of questions. Review your top pages quarterly, update the numbers, and change the last updated date only when you have made real changes.
Step 10: Track AI traffic and visibility
Most businesses have no idea how much traffic AI already sends them. Set this up in GA4: Admin, Data display, Channel groups, create a new group copied from the default, add a channel called AI Referral and place it above Referral, because order matters. Set the condition to Source matches regex:
chatgpt\.com|chat\.openai\.com|perplexity\.ai|gemini\.google\.com|copilot\.microsoft\.com|claude\.aiYou will now see AI visits, engagement and conversions as their own channel. Do not expect huge numbers. Conductor's benchmark report put AI referrals at roughly 1 percent of all website traffic, growing around 1 percent month on month, with ChatGPT driving 87.4 percent of it. That study covers May to September 2025 data across 13,770 domains, so treat it as a direction of travel rather than today's figure. The point is to track the trend and the conversion rate, which is usually well above average.
Source: Conductor AEO and GEO Benchmarks Report, data May to September 2025.
Note that AI Overview clicks show up as ordinary Google organic traffic, so they will not appear in this channel. Use your prompt audit to track those.
The KPIs that prove LLM SEO is working
| KPI | How to measure | Check |
|---|---|---|
| Prompt mention rate | Share of target prompts where you are named | Monthly |
| Citation rate | Share of prompts where you are linked as a source | Monthly |
| AI share of voice | Your mentions against total brand mentions across your prompts | Monthly |
| Accuracy and sentiment | Is the AI describing you correctly and positively | Monthly |
| AI referral sessions | GA4 AI Referral channel | Weekly |
| AI referral conversion rate | Leads divided by AI sessions | Monthly |
| Branded search volume | Search Console impressions for your brand name | Monthly |
Branded search is worth watching. People often discover a business in ChatGPT, then Google it by name to check. Rising branded impressions are frequently the clearest sign your AI visibility is growing.
What to expect, realistically
| Timeframe | Typical results |
|---|---|
| Weeks 1 to 4 | Crawler access fixed, Bing indexing live, baseline recorded, first AI referral data in GA4 |
| Months 2 to 3 | Restructured pages start being cited in live search engines |
| Months 3 to 6 | Third party mentions build, recommendation prompts start including you |
| 6 months plus | Consistent share of voice, with AI referrals a measurable lead source |
Anyone promising guaranteed ChatGPT rankings is overselling. AI answers vary and nobody controls them. What you can control is making your brand the most accessible, credible and widely corroborated option, which is what these ten steps do.
Frequently asked questions
Can you pay to rank in ChatGPT?
Not in the organic answers. ChatGPT's recommendations are not for sale, so the only route into them is being genuinely visible and credible across the web. Be sceptical of anyone selling guaranteed placement.
Why does ChatGPT recommend my competitors but not me?
Usually because your competitors have more third party mentions, more reviews or clearer content, or because your site blocks AI crawlers. Running a baseline prompt audit and checking which sources the AI cites will show you exactly where the gap is.
Does ranking on Google help me rank in ChatGPT?
Yes, indirectly. Strong SEO gets your pages indexed and trusted, which makes them available to AI tools that search the web. But ranking does not guarantee citation, so you also need extractable content and wider brand mentions.
How do I know if ChatGPT is sending me traffic?
Set up a custom AI Referral channel in GA4 using a regex that matches ChatGPT, Perplexity, Gemini, Copilot and Claude referrers. Some AI traffic arrives without a referrer, so rising branded searches are a useful secondary signal.
How often should I check my AI visibility?
Monthly is enough for most businesses. AI answers fluctuate day to day, so test each prompt several times per check and track the trend rather than any single result.
Where this leads
If this is the bit you want handled rather than explained:
- AI SEO. These ten steps run for you, with a baseline you can hold me to.
- GEO services. The citation side, tracked against a fixed question set.
- A free first pass. I will run your baseline prompt audit and send you what the assistants say.
