How small businesses get found by ChatGPT and AI assistants
By MarketWar OS ·
How small businesses get found by ChatGPT and AI assistants
ChatGPT and other AI assistants are increasingly used by customers to find local services, product recommendations and business details. Small businesses can optimise for "AI search" the same way they do for web search — but with a few critical differences.
Why this matters: AI assistants summarise and cite sources, often pulling from knowledge bases and LLM-indexed content. If you want to get cited by ChatGPT and other assistants, you need to make your information discoverable, reliable and machine-readable.
Understand what "AI search optimisation" means
AI search optimisation (AISO) is the practice of structuring your online presence so large language models (LLMs) and AI assistants can find, verify and reuse your content. It combines traditional SEO, local/GEO signals, structured data, authoritative citations and emerging conventions like llms.txt.
Key goals:
Be discoverable: ensure crawlers and ingestion pipelines can access your pages.
Be credible: present verifiable facts and official sources.
Be reusable: provide clear structured data and canonicalised content that assistants can extract.
Practical checklist to get cited by ChatGPT
1.Make your core facts crawlable and canonical
- Publish a clearly labelled "About", "Services", "Opening hours" and "Contact" pages.
- Use canonical tags and consistent business names across all platforms to avoid fragmentation.
- Avoid content behind JavaScript-heavy forms or sign-ins.
2.Use structured data (Schema.org)
- Implement LocalBusiness schema (or relevant subtype) for address, GEO coordinates, openingHours, telephone and priceRange.
- Include sameAs links to authoritative profiles (Google Business Profile, industry directories).
- LLMs favour explicit fields; structured data helps them extract facts directly.
3.Prioritise GEO and local signals
- Add precise GEO coordinates in your schema and visible footer.
- Maintain a Google Business Profile and keep NAP (name, address, phone) identical across directories.
- Collect local reviews on diverse platforms — AI assistants often weigh multiple independent sources.
4.Optimise for answerable content and snippets
- Create short, factual FAQ sections on each key page with concise Q&A (one question/one short answer).
- Use headings and plain language so LLMs can extract direct answers.
- Add step-by-step and “how to use” content for products or services; these are frequently quoted.
5.Make sources verifiable
- Where you state facts (hours, pricing, certifications), link to primary evidence: permits, certificates, PDF leaflets, press mentions.
- Encourage local press coverage and industry directory listings; third-party citations increase credibility.
6.Use llms.txt or a site-access policy
- llms.txt is an emerging convention (like robots.txt) to communicate to LLM builders which site content they may ingest or should ignore.
- If you want your content discoverable for reuse, create an llms.txt that explicitly permits public AI ingestion for relevant paths.
- If you prefer limited reuse, specify clear disallow rules. Note: adoption varies by provider, but using llms.txt is a forward-looking signal.
7.Serve concise machine-readable content
- Provide plain-text alternatives for complex pages (print-friendly or /text/ versions).
- Expose JSON-LD and downloadable data feeds for product catalogues or menus — these make ingestion reliable.
8.Build topical authority with focused content
- Publish localised service pages (e.g., “Window fitters in Hackney — same-day repairs”) rather than a single generic page.
- Link these pages internally and get backlinks from local organisations and chambers of commerce.
- Consistent topical coverage helps models associate your brand with specific queries.
9.Keep content fresh and correct
- Update opening hours for holidays and maintain change logs. AI assistants prefer current information and may avoid outdated sources.
- Use change dates in metadata and structured data to indicate freshness.
10.Monitor provenance and citations
- Track when your site is being cited by AI-driven answers using mentions, backlink alerts and AI-specific monitoring tools.
- If you find incorrect citations, correct the source content and issue clarifying pages or updates so assistants can re-index reliable information.
Extra tips for developers and technically-minded teams
Provide an API or open data endpoint for frequently requested items (price lists, availability). Well-documented APIs are easier for platforms to ingest.
Offer machine-readable sitemaps and an index of important URLs.
Use ALLOWED/NOINDEX headers deliberately: prevent low-quality or duplicate pages from polluting AI training data.
Consider contributing data to public knowledge graphs or Wikidata to increase the chance of appearing in knowledge panels.
Realistic expectations
AI assistants don’t “rank” websites like search engines; they synthesise answers from multiple sources. Getting cited by ChatGPT means being one of the clear, accurate, and accessible sources on a given topic. It’s rarely instantaneous—building consistent local signals, structured data and trustworthy citations takes time, but it pays off in higher visibility across AI search.
If you want practical help auditing your site for AI discoverability, MarketWar OS can run an AI search optimisation review and provide an actionable roadmap — book a free audit to get started.