Skip to content
NewNew guide. How to get your business into Google AI Overviews.Read it

Updated

Playbook · Search

The 30-day AI search visibility playbook

Four weeks, about three hours a week. Find out where AI answers name you today, clear what blocks them, and set up a monthly check you can keep.

OutcomeKnow where AI answers cite you, fix the blockers, track it monthly.

Time to run
4 weeks, ~3 hrs a week
Level
Intermediate
Who runs it
Founder, Marketer
Tools
  • Google AI Overviews
  • ChatGPT
  • Perplexity
  • Microsoft Copilot
  • Google Search Console
  • Bing Webmaster Tools
  • Google Analytics 4
  • Google Business Profile

To improve your AI search visibility, first log where Google's AI, ChatGPT and Perplexity name you for 20 buyer questions. Then let their search bots read your key pages, rewrite five pages answer-first with facts only you have, fix your profile, reviews and listings, and re-run the same questions every month.

AI assistants don't keep a list of businesses they like. They run searches, read the pages that come back, and quote the passages that answer best. So getting named is three plain jobs: be readable, be the clearest answer, and be mentioned by others. This is the practical side of AEO, and this playbook runs it in order over four weeks, with one person doing most of it in about three hours a week.

It is written for a business that already has a website and customers, such as a dental clinic, a school, a D2C skincare brand or a SaaS company, and wants to be named when buyers ask Google, ChatGPT or Perplexity for a shortlist. We run the same steps for clients in our AI search optimisation work.

Chapter 01: Baseline: where you appear today

Week 1

List 20 questions your buyers ask before they pick a business like yours, run each one in Google's AI features, ChatGPT and Perplexity, and log who gets cited and who gets named. This is your before picture. Answers change from run to run, so ask each question three times and look for the pattern, not the single answer.

Build the question list

Collect questions from places buyers already talk: enquiry forms, WhatsApp chats, sales and support calls, Google's "People also ask" box, and the follow-up questions the assistants suggest. Write them the way a buyer would type them, not the way your website describes you.

Then add the sub-questions. Google says AI Overviews and AI Mode may use "query fan-out", running several related searches across subtopics before they answer. Ahrefs found that by March 2026 only about 38% of pages cited in AI Overviews ranked in the top ten for the original question, and it links the drop to fan-out. A clinic's "best clear aligners in Chennai" fans out into cost, treatment time, who does the treatment and reviews. A school's "good CBSE school in Whitefield" fans out into fees, admission dates and transport.

Run and log

Run every question in Google (AI Overviews and AI Mode), ChatGPT with search on, and Perplexity. Record two things separately: were you named in the answer, and was your page cited as a source? They are different results and move for different reasons. BrightLocal found only 20 to 33% average overlap between the results of repeat runs of the same local prompt, so three runs per question is the minimum that tells you anything.

Worksheet Baseline log
QuestionEngineRunNamed?Cited?Who was named or cited instead
Best clear aligners clinic in Anna NagarGoogle AI Mode1NoNoTwo clinics, one directory
Best clear aligners clinic in Anna NagarChatGPT1blankblankblank
blankblankblankblankblankblank
blankblankblankblankblankblank
One row per run. Three runs per question per engine. Keep this sheet: you will fill it again in week 4.
You'll have
  • A baseline sheet with 20 questions, three runs each, across three engines
  • A short list of who wins the questions you want to win

Chapter 02: Access: make sure the bots can read you

Week 1

Check that your key pages are indexed in Google and Bing, are not set to noindex or nosnippet, and that robots.txt lets the search bots in: Googlebot, Bingbot, OAI-SearchBot, Claude-SearchBot and PerplexityBot. Training bots are a separate choice. Blocking a search bot can remove you from that engine's answers.

Read your robots.txt

Open yoursite.com/robots.txt. A Disallow: / under any of those names, or under User-agent: *, is the first fix. OpenAI says OAI-SearchBot is what surfaces sites in ChatGPT search, while GPTBot is for training. Anthropic says blocking Claude-SearchBot "may reduce a site's visibility" in Claude's search. Perplexity says PerplexityBot is designed to surface and link websites in its results and is not used to crawl content for AI foundation models. So you can keep training bots blocked and still let search in:

User-agent: OAI-SearchBot
Allow: /

User-agent: Claude-SearchBot
Allow: /

User-agent: PerplexityBot
Allow: /

User-agent: Bingbot
Allow: /

Then check the layer above. A firewall or CDN bot rule can return an error to a crawler before it ever reads robots.txt. Look for those names in your hosting or CDN's bot settings, and in your server logs if you can.

Get indexed in Google and Bing

Google says there is no special file, schema or markup needed for its AI features: a page has to be indexed and allowed to show a snippet. In Google Search Console, use URL Inspection on your ten most important pages and remove any accidental noindex or nosnippet.

Bing matters more than most businesses think, because Microsoft Copilot answers from it. Verify your site in Bing Webmaster Tools, submit your sitemap, and switch on IndexNow, the protocol that tells Bing and other engines the moment a page changes. Many CMS plugins and CDNs offer it as a setting.

Worksheet Access table
PageIndexed in Google?Indexed in Bing?noindex / nosnippet?Search bots allowed?
/aligners/YesNoNoneYes
blankblankblankblankblank
blankblankblankblankblank
Pass or fail per page. Fix every fail before you rewrite anything.
You'll have
  • A pass or fail access table for your ten most important pages
  • Bing Webmaster Tools verified, sitemap submitted, IndexNow on

Chapter 03: Answers: rewrite your top pages

Weeks 2–3

Pick the five pages that should win your baseline questions. Give every section a direct answer in its first sentences, name things exactly, add specifics only you have, show who wrote it and when it was last updated, and cite your sources. Don't chop pages into tiny fragments or stuff them with keywords.

Put the answer first

Assistants quote passages, so each answer has to stand on its own. Put the question in a heading and answer it in the first 40 to 60 words beneath it. Position matters too: Kevin Indig's analysis of 18,012 ChatGPT citations found 44.2% came from the first 30% of a page.

Weak: "Every smile is unique, and our experienced team will create a plan tailored to you."

Lifts cleanly: "Clear aligners at our Anna Nagar clinic take 6 to 18 months for most adults. We confirm the plan after a 3D scan, which is part of the first consultation."

The second names the place, the range, the method and the next step. Do the same for prices, service areas, timings and eligibility: plain text on the page, not a PDF or an image. Our guide on how to appear in Google AI Overviews goes deeper on Google's side of this.

Add what only you know

Look at the five pages that win your questions today and ask what yours has that theirs don't. Usually it's first-hand detail: how you actually run the treatment, the admission timeline your school really follows, the ingredient test your brand did. In the GEO study from Princeton and IIT Delhi, adding statistics, quotations and cited sources made pages more visible in generated answers in lab tests. Expect smaller effects on live engines, but the habit is free.

Show who wrote it and when

This is what Google's E-E-A-T guidance asks for in practice. Add an author and a reviewer, show a real updated date, and change that date only when the content changes. Ahrefs found cited pages were fresher than ordinary search results on average, which is exactly why date-bumping tempts people. A new date on old content earns nothing.

The same rewrite tends to help ordinary rankings too, which is what our SEO work is built on. Running an online store? The product page SEO playbook applies the same idea to product pages.

You'll have
  • Five rewritten pages, each answering its questions in the first lines
  • Sources cited and a real updated date on each

Chapter 04: Mentions: be findable beyond your site

Week 3

AI engines lean on pages other than yours. Complete your Google Business Profile, collect honest reviews, make sure your name, address and description match on the directories your buyers use, and show up where your category is discussed. In Ahrefs' data, brand mentions correlate with AI Overview visibility more strongly than links do.

Profile, reviews and listings

In Ahrefs' study of 75,000 brands, mentions of a brand across the web correlated with AI Overview visibility at 0.664, against 0.218 for backlinks. That's a correlation, not proof, but it points to where the work is. For local questions, BrightLocal's 2026 study of about 1.9 million AI citations found Google Business Profile was the largest single source, at about 29%, with business websites at 42% overall and Yelp the next-largest single source.

So for a clinic or a school: complete every field of the profile, keep hours and services current, and make asking for a review part of the service, not a campaign. Then line up your name, address and phone on the five listings your buyers actually use. Our local SEO team runs this for multi-location clinics and schools, and the Google Business Profile ranking factors guide covers what moves a profile.

Show up where buyers ask

Find the forums, groups and videos where your buyers ask the questions on your list, and answer one properly, as a person, with something useful. A D2C brand can do the same with honest reviewers and creators. Earn a place on other people's lists rather than publishing your own "best of" list with yourself on top.

You'll have
  • A list of 10 third-party places, each updated or claimed
  • A review request built into how you deliver the service

Chapter 05: Measure: set up the monthly habit

Week 4

Re-run the baseline questions, compare them with week 1, and set a monthly routine: your 20-question log, AI-assistant referrals in GA4, and the AI reports Google and Microsoft now offer. Judge direction over months, not days, because AI answers are volatile and no single tool covers every engine.

Re-run and compare

Run the same 20 questions, three times each, in the same engines. Count the share of runs where you were named and the share where you were cited. Compare with week 1 and write down what changed on your side in between, so you can connect cause and effect over time.

Read the first-party reports

Bing Webmaster Tools' AI Performance report shows how often your pages are cited in Microsoft Copilot and Bing's AI summaries, which pages, and the grounding queries behind them. It counts citations, not clicks. Google's Search generative AI performance reports in Search Console show impressions in its AI features, and sites with very few AI impressions may not see them yet. Neither covers ChatGPT, Perplexity or Claude, which is why your question log stays.

In GA4, build one exploration that filters sessions from chatgpt.com, perplexity.ai, copilot.microsoft.com and gemini.google.com. ChatGPT often tags links with utm_source=chatgpt.com (a UTM parameter), but some visits arrive with no referrer and land in Direct, so read the number as a floor. And keep the claim honest: a citation is not a click, and a click is not a customer. To see whether those visits turn into enquiries, set up key events first; our GA4 conversion tracking guide walks through it.

You'll have
  • A dated day-30 comparison against week 1
  • A monthly 30-minute review in the calendar, with the same 20 questions
Common traps

Where this usually goes wrong.

Each one costs a month. Skip them and the run gets shorter.

Before you start

Questions people ask first.

Access fixes are quick: OpenAI says a robots.txt change for OAI-SearchBot takes about 24 hours to apply. Being cited takes longer, because assistants pick pages that rank for the questions they search and brands other sites mention. Expect early movement within weeks and judge it on a monthly baseline, not a single answer.

No. Google says Search doesn't use llms.txt or other AI text files, and neither OpenAI nor Perplexity documents it as a way to appear in answers. It costs little to keep, so keep it if you have one. Spend the time on readable answers and outside mentions instead. More in our llms.txt entry.

That is a business choice about training, separate from search. OpenAI says GPTBot collects training data while OAI-SearchBot decides whether your site appears in ChatGPT search. Anthropic splits ClaudeBot (training) from Claude-SearchBot (search) the same way. You can block the training bots and keep the search bots allowed.

Not in the organic answer. Google says there is no special requirement or markup to appear in AI Overviews or AI Mode, and its spam policies now cover attempts to manipulate generative AI responses. Anyone selling a guaranteed placement is selling something the engines say does not exist.

Seer Interactive reports that ChatGPT often adds utm_source=chatgpt.com to the links it sends, so build a GA4 exploration filtered on that source plus perplexity.ai, copilot.microsoft.com and gemini.google.com. Some visits arrive without a referrer and show as Direct, so treat the number as a floor, not the full count.

Sources and method

Sunandha ThamraislevanFounder & CEO, BridgeE

Founder and CEO of BridgeE, a digital marketing agency in Chennai. She leads the strategy for every client and agrees the number each plan is judged on before the work starts.

How we wrote thisFrom the search engines' and AI companies' own documentation and the studies listed here. Every number in this playbook comes from one of them, with its date. Checked 1 Nov 2026.

  1. AI features and your websiteGoogle Search Central · December 2025
  2. Google's guide to optimizing for generative AI features on Google SearchGoogle Search Central · July 2026
  3. Overview of OpenAI crawlersOpenAI · 2026
  4. Does Anthropic crawl data from the web, and how can site owners block the crawler?Anthropic Support · April 2026
  5. Perplexity crawlersPerplexity · 2026
  6. Introducing AI Performance in Bing Webmaster Tools (public preview)Microsoft Bing · February 2026
  7. IndexNowIndexNow.org · 2026
  8. AI Overview citations from the top 10Ahrefs · March 2026
  9. AI assistants prefer to cite fresher contentAhrefs · July 2025
  10. AI Overview brand visibility factors (75K brands)Ahrefs · May 2025
  11. Where ChatGPT citations come from on the pageSearch Engine Land · February 2026
  12. Local AI visibility studyBrightLocal · September 2026
  13. Search Generative AI performance reports in Search ConsoleGoogle Search Central · June 2026
  14. Generative AI performance reportSearch Console Help · 2026
  15. Spam policies for Google web searchGoogle Search Central · August 2026
  16. Are AI sites like ChatGPT sending your website traffic?Seer Interactive · June 2026
  17. GEO: Generative Engine OptimizationarXiv (Aggarwal et al., KDD 2024) · June 2024

Bring your number. We'll bring the plan.

A 30-minute call about AI search (AEO & GEO). You leave with three things to fix first, whether you hire us or not.

Book a call