Playbook · AI Search
How to get found in ChatGPT, Perplexity and AI search results
Ranking in Google and being cited by an AI model are related but not the same game.
- 9 ranked moves
- 4-week sequence
- For teams with an existing content library
How to get found in ChatGPT, Perplexity and AI search results: what should you actually do?
AI answer engines pull from pages that state facts plainly, cite sources, and answer a question in the first two sentences. Start by publishing clear, structured answers to the 20 questions your buyers actually ask, get cited on a handful of sites the models already trust, and check monthly whether your brand shows up in test prompts. This takes 8-12 weeks to show movement, not days.
Last updated 3 August 2026
| Play | Effort | Cost |
|---|---|---|
| Test your current AI visibility before changing anything | 1 day | $0 |
| Rewrite your best pages to answer the question in the first two sentences | 2-3 days | $0-$1,500 if outsourcing the edit |
| Add specific numbers and named sources to every claim | 2-3 days | $0 |
| Get cited on sites the models already trust | Ongoing, 2-3 hours/week | $0-$300/month for a PR tool |
| Build a structured FAQ block on every key page | 1-2 days | $0 |
| Add FAQ and Article schema markup | 1 day | $0 |
| Publish one original data asset a quarter | 3-4 weeks | $1,500-$6,000 |
| Keep your Google Business Profile and review sites current | 2-3 hours | $0 |
| Re-test visibility monthly and track which pages get cited | 2 hours/month | $0 |
AI answer engines like ChatGPT, Perplexity and Google's AI Overviews don't crawl and rank pages the way traditional search does. They synthesise an answer from a smaller set of sources they judge trustworthy and well-structured, often pulling from third-party mentions of your brand as much as your own site. That means visibility here depends on being cited elsewhere, not just publishing on your own domain.
This plan is for a company with an existing content library and at least a basic understanding of its buyer questions. It won't help if you have no content at all — there's nothing for a model to find or cite yet.
- — A B2B or B2C brand with at least 15-20 published pages of real substance
- — A team willing to publish plainly stated facts and numbers, not vague claims
- — Some existing press mentions, reviews, or third-party citations to build from
- — Patience for an 8-12 week feedback loop rather than a weekly ranking check
If you have no published content and no third-party mentions anywhere, start with basic SEO and PR first. AI visibility largely rides on top of an existing footprint — it doesn't create one from nothing.
The moves
Ranked, highest return first.
Work down the list. Each one names the first step so there's nothing to plan.
- 01
Test your current AI visibility before changing anything
You need a baseline to know if any of this is working. Most teams have never actually asked an AI model about their own category.
First step: Write 15-20 real buyer questions and run each through ChatGPT, Perplexity and Google's AI Overview, logging whether your brand appears and what's cited instead.
- Tools
- ChatGPT, Perplexity, a spreadsheet
- Effort
- 1 day
- Cost
- $0
- 02
Rewrite your best pages to answer the question in the first two sentences
Models extract answers from the clearest, most direct statement on a page. Pages that bury the answer under three paragraphs of scene-setting get skipped.
First step: Pick your 10 highest-value pages and rewrite the opening so the direct answer to the page's core question appears in the first 40 words.
- Tools
- Your CMS, Google Docs
- Effort
- 2-3 days
- Cost
- $0-$1,500 if outsourcing the edit
- 03
Add specific numbers and named sources to every claim
Vague claims like 'many companies see improvement' get filtered out. A stated figure with a source attached gets picked up and quoted directly.
First step: Audit your top pages for any unsupported claim and replace it with a specific number, date, or named study.
- Tools
- Your existing data, published research
- Effort
- 2-3 days
- Cost
- $0
- 04
Get cited on sites the models already trust
AI models weight citations from sites with an established trust signal — review platforms, trade press, Wikipedia-adjacent sources — more heavily than a brand's own blog.
First step: List the 10 sites your competitors get quoted or reviewed on, and pitch a comment, data point or guest contribution to three of them this month.
- Tools
- HARO/Qwoted, a media list, email
- Effort
- Ongoing, 2-3 hours/week
- Cost
- $0-$300/month for a PR tool
- 05
Build a structured FAQ block on every key page
Question-and-answer formatting is easier for a model to extract cleanly than a long paragraph, and it maps directly onto how people phrase prompts.
First step: Add a 4-6 question FAQ block to your top 10 pages using the exact phrasing your buyers use, not internal jargon.
- Tools
- Your CMS, schema markup plugin
- Effort
- 1-2 days
- Cost
- $0
- 06
Add FAQ and Article schema markup
Structured data doesn't guarantee a citation, but it removes ambiguity about what a page is answering, which helps both traditional and AI crawlers parse it correctly.
First step: Add FAQPage and Article schema to your top 20 pages and validate with a schema testing tool.
- Tools
- Google's Rich Results Test, Schema.org markup
- Effort
- 1 day
- Cost
- $0
- 07
Publish one original data asset a quarter
Models favour primary sources over secondary summaries. A survey or benchmark with your own numbers becomes something other content — and models — cite back to you.
First step: Pick one data set you can credibly gather (customer survey, product usage stats, pricing analysis) and publish it with a clear methodology section.
- Tools
- Typeform, a spreadsheet, a writer
- Effort
- 3-4 weeks
- Cost
- $1,500-$6,000
- 08
Keep your Google Business Profile and review sites current
Several AI tools pull location, pricing and review data from these profiles directly, independent of your own website.
First step: Update your Google Business Profile, G2 and Capterra listings with current pricing, features and a response to your three most recent reviews.
- Tools
- Google Business Profile, G2, Capterra
- Effort
- 2-3 hours
- Cost
- $0
- 09
Re-test visibility monthly and track which pages get cited
AI models update their training and retrieval sources on their own schedule. A monthly check tells you what's working without over-reacting to daily noise.
First step: Re-run your original 15-20 test prompts each month and log whether citations changed, adding new prompts as your content expands.
- Tools
- ChatGPT, Perplexity, a spreadsheet
- Effort
- 2 hours/month
- Cost
- $0
Sequence
What to do first, week by week.
Baseline and audit
Test current AI visibility across 15-20 real buyer questions and audit your top 10 pages for vague claims and buried answers.
Rewrite and structure
Rewrite openings to answer directly, add FAQ blocks and schema markup, and replace vague claims with specific sourced numbers.
Earn third-party citations
Pitch trade press and review platforms, and update your Google Business Profile, G2 and Capterra listings.
Publish original data and re-test
Kick off one original data asset for the quarter and re-run your test prompts to compare against the baseline.
Avoid
Where this usually goes wrong.
Chasing AI visibility with zero underlying content
There's nothing for a model to cite if you haven't published substantive pages yet. Build the content library first, then optimise how it's structured.
Treating this like a one-week SEO sprint
Model retrieval sources update slowly. Expect 8-12 weeks before any test-prompt movement, and don't panic-rewrite everything after one flat monthly check.
Ignoring third-party mentions entirely
A brand's own site is only part of the picture. Reviews, press and forum mentions often carry more weight in what a model decides to cite.
Questions
Common questions.
How is AI search optimisation different from SEO?
SEO ranks whole pages in a list; AI search synthesises a direct answer from a small set of trusted sources, often including third-party mentions. The tactics overlap (clear structure, real data) but the goal shifts from ranking to being the source a model actually quotes.
How long before we see results?
Most teams see the first visibility changes in test prompts after 8-12 weeks, since model retrieval and training sources don't update in real time. Track monthly rather than daily to avoid over-reacting to noise.
Do we need schema markup for this to work?
It helps but isn't the deciding factor. Schema removes ambiguity about what a page answers, but a model still needs plain, well-sourced answer text to actually cite. Prioritise the writing first, schema second.
Can we track whether ChatGPT is sending us traffic?
Some analytics tools now show referral traffic tagged from chatgpt.com and perplexity.ai as a source, though volume is usually small compared to organic search today. Track it as a leading indicator, not yet a primary channel.
More playbooks
Other plans worth reading.
Partner and integration growth plan for B2B SaaS
ReadPlaybook · ActivationSaaS onboarding and activation plan to fix a weak trial-to-paid rate
ReadPlaybook · Win-backWin-back plan for lapsed and churned B2B SaaS customers
ReadPlaybook · App retentionRetention plan for a B2C app with high day-1 or day-7 drop-off
ReadPlaybook · B2B EcommerceGrowth plan for wholesale and B2B ecommerce sellers
ReadPlaybook · ReferralsReferral engine for professional services firms
ReadElsewhere on gtm.help
Keep exploring.
Tool guides
What to use, what it costs, when it earns its keep.
BrowseTool comparisons
Head-to-head picks for the calls people get stuck on.
BrowseMarketing checklists
Task lists by business type and competitive situation.
BrowseStraight answers
The growth questions founders ask, answered with numbers.
BrowseGet started
Get this ranked for your business.
Paste your URL and you'll have a plan built from your site and competitors in about a minute.