How to Show Up in ChatGPT and Perplexity

When a homeowner asks ChatGPT "which roofer do you recommend in Manchester," the model does not query a single database or ring your office. It synthesises a response from multiple public sources: your website, your directory listings, your reviews, and editorial mentions of your business across the web.

This guide breaks down how each major AI engine retrieves local business information, which signals carry the most weight, and which concrete actions move the needle for home services and construction businesses in the UK.

If you are new to the underlying concept, the article on what GEO means for trades covers the strategic framework. This piece focuses on the operational mechanics.

---

Why AI Search Engines Work Differently from Google

Google returns a ranked list of URLs. ChatGPT, Perplexity, and Google AI Overviews generate a synthesised answer that may include your business name, address, rating, and a description of your services — all in one paragraph, without the user ever visiting your website.

That changes the logic of visibility in three important ways:

• Ranking in Google is not enough. ChatGPT retrieves information primarily through the Bing index, not Google's. A business invisible on Bing may rank well on Google but never appear in ChatGPT responses.
• Google Reviews do not reach ChatGPT. ChatGPT cannot access Google Business Profile. It reads Yelp, Foursquare, the Better Business Bureau, and vertical directories.
• 86% of AI citations come from content the brand controls directly — its own website, directory profiles, and review platform listings — according to Yext's analysis of 6.8 million AI citations.

The mechanism driving this is retrieval-augmented generation (RAG). Rather than relying solely on knowledge baked in during training, the AI system fetches external documents at query time and grounds its answer in those sources. Peer-reviewed research confirms that grounding LLMs in retrieved passages substantially reduces hallucinated content and increases citation accuracy (AGREE framework, arXiv:2311.09533; Citation-Enhanced Generation, arXiv:2402.16063). For your business, this means the quality and accessibility of your web presence determines whether the model cites you — not just your Google ranking.

---

Which Sources Each Engine Consults

The table below summarises findings from BrightLocal (800 local ChatGPT queries), Yext (6.8 million citations), Cheers GEO Academy, and Soci.ai.

AI Engine	Primary source	Key directories	Reads Google Reviews?
ChatGPT	Bing index + own website (58% of citations)	Yelp, Foursquare, BBB, Three Best Rated	No
Perplexity	Vertical directories + Yelp (data agreement)	Checkatrade, Houzz, Rated People, TripAdvisor	No
Google AI Overviews	Google index + Google Business Profile	TripAdvisor, own website	Yes (via GBP)
Microsoft Copilot	Bing index (same channel as ChatGPT)	Yelp, TripAdvisor, Bing Places	No
Gemini (standalone)	Own website + JSON-LD schema (52% of citations)	Google Maps, GBP direct	Yes (direct)

---

The Five Sources AI Engines Read About Your Business

1. Your Own Website

Your website is the most cited source: 58% of ChatGPT references point directly to the business's own domain, according to BrightLocal's analysis of 800 local queries. AI crawlers read the initial HTML response. If your website renders content exclusively via client-side JavaScript without server-side rendering (SSR), the bots from OpenAI, Anthropic, and Perplexity cannot parse it.

What to implement:

• Individual service pages with detailed descriptions, service area coverage, and indicative pricing.
• schema.org/LocalBusiness markup with name, address, phone, hours, services, and aggregate rating.
• FAQ sections with FAQPage schema on high-traffic pages.
• SSR or static site generation (SSG): content must appear in the server-delivered HTML, not only after JavaScript executes in the browser.

For a deeper look at the visibility framework, see the visibility pillar and the GEO glossary.

2. Directory Listings and NAP Consistency

NAP — Name, Address, Phone — is the canonical term for a local business's identity data across the web. When ChatGPT finds inconsistent data between your website, your Bing Places profile, your Yelp listing, and other directories, it reduces confidence in your business and may omit it from responses entirely.

Directory weight by engine:

• ChatGPT / Copilot: Yelp, Foursquare, BBB, Three Best Rated, Expertise.com, Thumbtack.
• Perplexity: Checkatrade, Rated People, Houzz, and trade-specific vertical directories.
• Google AI Overviews / Gemini: Google Business Profile, TripAdvisor, and the business's own website.

Consistency matters more than volume. A business with identical NAP data in fifteen directories outperforms one with perfect listings in three but conflicting data everywhere else. Audit your listings before expanding to new directories.

3. Reviews and Ratings

Reviews are not only social proof for human visitors. AI engines read them to infer quality, reliability, and the specific services a business provides.

• ChatGPT reads Yelp, Foursquare, and BBB reviews. It cannot read Google Reviews directly, but does encounter them when another page quotes or aggregates them.
• Gemini and Google AI Overviews prioritise Google Business Profile ratings.
• Perplexity extracts scores from Yelp and vertical directories, with Yelp appearing as the first or second source in most local responses.

Review text matters as much as star count. Models analyse sentiment and extract mentions of specific services. A review stating "fitted the boiler in four hours without leaving a mess" provides more actionable signal than "great company, five stars." Specificity is what gets extracted, synthesised, and cited in AI responses.

For the research behind how reviews influence trust and AI citation behaviour, the article on the science of online reviews covers the evidence in depth.

4. Press Mentions and Editorial Content

27% of ChatGPT's sources are third-party business mentions — local press, editorial listicles of the "best tradespeople in [city]" format, and trade publications — according to BrightLocal. For informational queries such as "how much does a bathroom renovation cost," blog articles account for up to 86% of Perplexity citations.

What generates citable editorial coverage:

• Mentions in local media and trade publications with your business name, location, and speciality clearly stated.
• Inclusion in "best trades in [city] 2026" listicles on credible local domains.
• Publishing your own content with cited data and statistics. Research by Soci.ai found that content containing quantified claims with named sources receives measurably higher AI citation rates than qualitative-only content.

A citation from a credible local publication carries more weight than ten blog posts on your own domain, because it represents an independent third party affirming your business's existence and quality — the kind of corroboration AI models are designed to treat as a higher-confidence signal.

5. Structured Data (schema.org)

Schema markup translates your website content into a format AI models can read without ambiguity. Gemini uses it as a primary source: 52% of its citations come from websites with correct LocalBusiness schema, compared to 15% from those without it, according to Yext.

The most impactful schema types for local home services businesses:

• LocalBusiness or trade-specific subtypes: Plumber, GeneralContractor, RoofingContractor, ElectricalContractor
• FAQPage for question-and-answer sections
• Review and AggregateRating for customer ratings
• Service for each individual offering

A generic, incompletely populated schema yields a 41.6% AI citation rate. A specific, complete implementation raises that to 61.7%, according to Fuelonline's analysis of AI crawler behaviour. That gap is large enough to justify prioritising schema completion over most other technical improvements.

---

Do Not Block AI Retrieval Bots

Many WordPress plugins, Cloudflare WAF configurations, and CDN rule sets block AI bots automatically — often without distinguishing between bot types. This is one of the most consequential and least visible mistakes a local business can make.

There are two distinct categories:

• Training crawlers (GPTBot, ClaudeBot, Google-Extended): collect data to train AI models. You can block these without affecting your visibility in AI-generated answers.
• Real-time retrieval bots (OAI-SearchBot, PerplexityBot, Claude-SearchBot): index or retrieve content to generate responses for users right now. Blocking these makes your business invisible on those platforms.

Research by BuzzStream found that 71% of major publishers block at least one retrieval bot, many unintentionally by enabling a blanket "block all AI bots" option in their SEO plugin or Cloudflare dashboard. The consequence is direct: your content becomes ineligible for citation regardless of how strong your NAP consistency or review volume may be.

For more on this anti-pattern and how to avoid it, see the visibility section.

Check your robots.txt file at yourdomain.co.uk/robots.txt and verify that the following agents have explicit access:

User-agent: OAI-SearchBot
Allow: /

User-agent: PerplexityBot
Allow: /

User-agent: Claude-SearchBot
Allow: /

If you use Cloudflare or another WAF, check network-level firewall rules as well. They can block bots before robots.txt is ever read.

---

IndexNow: Faster Bing Indexing Means Faster ChatGPT Visibility

ChatGPT retrieves information through the Bing index. Without IndexNow, Bing can take one to four weeks to rediscover an updated page. With IndexNow, that window narrows to three to seven days.

The protocol is free and supported by Bing, Yandex, Naver, and Seznam. Google does not support it; Google Search Console remains the required mechanism for Google AI Overviews and Gemini.

In 2026, 80 million websites use IndexNow with five billion URLs submitted daily. Every time you update a service page or publish a blog article, a competitor with IndexNow active reaches the Bing index sooner — and therefore appears in ChatGPT responses before you do.

Implementation is straightforward: generate an API key, host it at your root domain, and ping the IndexNow endpoint each time content changes. Most modern CMS platforms and Next.js deployments support this through a plugin or a short API call.

---

How to Measure Your AI Engine Presence

No centralised panel exists for cross-engine AI citation tracking in 2026. Practical options:

• Manual queries: search "[your trade] in [your city]" in ChatGPT, Perplexity, and Gemini. Note whether you appear, what information is cited, and which competitors are included.
• Bing Webmaster Tools: added an AI performance panel in February 2026 showing when your content is cited. Bing acknowledges that 99.6% of AI content usage is invisible to publishers, so data is partial by design.
• Specialist tools: BrightLocal, Cheers, Cited, and Soci.ai offer AI mention tracking with varying coverage and pricing. See the comparisons section for a breakdown of options.

Run manual queries monthly, track results in a spreadsheet, and compare against two or three direct competitors. The gaps that surface — missing directories, no Yelp reviews, blocked retrieval bots — give you a prioritised action list.

---

Priority Checklist

Ordered by estimated citation impact based on the research cited throughout this article.

High priority

• Website delivers content in server-rendered HTML (SSR or SSG), not only client-side JavaScript.
• Complete LocalBusiness schema with name, address, phone, hours, services, and aggregate rating on every relevant page.
• Bing Places profile claimed, verified, and kept current.
• Yelp and Checkatrade profiles optimised with description, photos, service categories, and active review management.
• Identical NAP data across your website, Google Business Profile, Bing Places, Yelp, and all major directories.
• robots.txt explicitly allows OAI-SearchBot, PerplexityBot, and Claude-SearchBot.

Medium priority

• IndexNow implemented to reduce Bing re-indexing time after each update.
• Individual service pages per speciality and geographic area served.
• Active review acquisition strategy on Yelp and trade-specific vertical directories in addition to Google.
• Presence in at least two or three vertical directories relevant to your specific trade.

Complementary

• Appearances in local editorial listicles and trade publications.
• Blog content with data, statistics, and named sources.
• FAQPage schema on high-traffic pages.
• Monthly manual citation monitoring across ChatGPT, Perplexity, and Gemini.

---

For the traditional local search signals that also feed Google AI Overviews, the article on how to rank on Google Maps for trades covers the parallel discipline in detail. Explore the blog for additional guides on structured data, review strategy, and GEO applied to construction and home services businesses across the UK.