How to get your business noticed in AI search

Step 1: Let the answer crawlers in, and get indexed in Google and Bing

The first and most commonly overlooked step is ensuring the crawlers that feed AI answers can actually reach your content, and that your site is in the indexes those AI assistants read.

What changed: until September 2026, ChatGPT search leaned heavily on Bing's index. It now runs its own. Peec AI reported on 4 September 2026 that ChatGPT searches an index built by OAI-SearchBot, with Bing used mainly in Deep Research. Bing still feeds Microsoft Copilot, and Google's index feeds AI Overviews, AI Mode and Gemini. So verify your site in both Google Search Console and Bing Webmaster Tools, and let OAI-SearchBot in for ChatGPT.

Open your robots.txt file (available at yourdomain.ie/robots.txt) and check how it treats the following crawlers. Some fetch pages for answers; others only collect training data, and blocking those costs you no answer visibility:

• GPTBot: OpenAI's training crawler. It gathers public web content used to train OpenAI's models; blocking it opts you out of training, not out of ChatGPT search
• OAI-SearchBot: OpenAI's search crawler. It indexes your pages so they can appear and be linked in ChatGPT search results
• ChatGPT-User: fetches a page live while ChatGPT writes an answer. Cloudflare classes it as an Agent, which it blocks by default on ad-bearing pages for sites that joined after 15 September 2026
• ClaudeBot: Anthropic's training crawler for Claude. Anthropic also runs Claude-SearchBot (search indexing) and Claude-User (live visits during conversations)
• Google-Extended: a robots.txt control rather than a separate crawler. It covers Gemini training and Gemini grounding together, so blocking it removes you from Gemini answers. It does not affect Search, AI Overviews or AI Mode, which use the standard Googlebot
• PerplexityBot: builds the search index for Perplexity. Perplexity says it is not used to train AI models
• Bingbot: used by Microsoft Bing, which powers Copilot and Bing's AI summaries

Your robots.txt should allow the answer crawlers. Whether to allow the training crawlers is a policy choice. Note that a crawler named in its own group ignores the rules under User-agent: *, so repeat any disallows there. A sound file looks like this:

Step 2: Structure your content for AI extraction

The way AI systems read and extract content is fundamentally different from how a human reader does. A human reads linearly and can follow an argument that builds over several paragraphs. An AI system scanning for citation-worthy content looks for the most direct, extractable answer, usually in the opening lines of a page or section.

This means the most important structural change most businesses need to make is moving the answer to the top. Instead of beginning a service page with "At [Your Business], we pride ourselves on delivering exceptional [service]…", begin with a direct answer to the question a customer would ask:

The second version answers the questions a customer would actually ask an AI: who you are, what you do, where you are, and how to contact you. It is extractable in a way the first version is not.

Content structure checklist for AI extraction

Apply this to every key page on your site:

• Direct answer or clear description in the first paragraph (no vague preambles)
• Use h2 and h3 headings that match real questions customers ask ("How does [service] work?", "What does [service] include?")
• Include a visible question-and-answer section on every service page with five to eight specific questions
• State your location, opening hours, and contact details clearly on every page
• Include specific details: prices, turnaround times, qualifications, certifications
• End each section with a clear next step, not a vague call to action

Step 3: Keep schema markup accurate, as hygiene

Schema markup is machine-readable code (added to your pages in JSON-LD format) that describes what your content is and who it belongs to. Until 2026 it was widely sold as an AI citation lever. The evidence now says otherwise: Ahrefs tracked 1,885 pages that added JSON-LD (May 2026) and found no measurable effect on AI citations, and Google says no special schema is needed for its AI features. Keep it because it is cheap and supports classic search features, not because it gets you cited.

For local businesses, the schema worth keeping accurate is:

• LocalBusiness schema: your business name, address, phone number, opening hours and category, matching your Google Business Profile exactly.
• FAQPage schema: optional. AI assistants quote the visible questions and answers; there is no independent evidence that the markup itself adds citations.
• Article schema: for guide pages, the author, publisher and honest published and updated dates.
• Service schema: describes individual services with their names, descriptions, and providers.

If your website runs on WordPress, plugins such as Yoast SEO or Rank Math can add basic schema automatically. For more precise control, JSON-LD can be added directly to your page's <head> section. Whatever you add must match the visible page. Our AI Foundation & SEO Build service covers it as part of the technical hygiene.

Step 4: Build answer-focused content pages

Beyond restructuring existing pages, the businesses that consistently appear in AI citations are those that publish content specifically designed to answer the questions their customers ask. AI search queries are long and specific: a 2026 SOCi study found the average AI assistant query is about 23 words, roughly six times longer than a typical search query. Each of those long queries represents a specific question someone is asking, and each question is an opportunity to be the cited source. Cover the five kinds of question buyers ask: what is it, how does it compare, what are the alternatives, will it suit my case, and which should I buy.

Consider what questions your customers ask before choosing your type of business, not just at the point of purchase. A restaurant owner might publish:

• "What to look for in a wedding venue restaurant in Galway"
• "Best restaurants for business lunches in Dublin city centre"
• "Family-friendly Sunday lunch options in Cork with high chairs"

A solicitor might publish:

• "What happens to a house when you get divorced in Ireland"
• "How long does probate take in Ireland"
• "What is the difference between a will and a living will in Irish law"

Each of these pages targets a real query that someone is asking AI right now. If your page answers it directly and authoritatively, it has a genuine chance of being cited. For more on what content to publish, see our AI Answer Content service.

Step 5: Establish authority signals

AI systems are designed to avoid citing low-quality, untrustworthy or anonymous sources. These signals help your own site look like a credible source, though the strongest authority comes from other sites (step 6):

• Organisation identity: clear About page with your business name, registration details, address, and founding information
• Author identity: real named authors on guide pages, linked to a short author page with credentials. The evidence is weak but positive; never invent an author
• Mentions elsewhere: other reputable websites in your sector or location naming you, with or without a link
• Maintained content: honest published and "last updated" dates on key pages, and real updates behind them
• Consistent named entities: your business name, location, and sector should appear consistently across your site, Google Business Profile, and any directories
• Content policies: a privacy policy, terms of service, and editorial standards page all signal that your site is operated by a credible organisation

Step 6: Build off-site authority (the most overlooked step)

This is where most businesses (and most AI SEO advice) fall short. In practice, most AI citations come from third-party sources like reviews, directories, roundups and press, not from brand-owned websites. Muck Rack's May 2026 analysis of 25 million AI citations found 84% were earned media; press releases were 1.1% and paid placements 0.3%.

This means a perfectly structured website with zero external presence will still struggle to be cited. The off-site authority steps that directly affect AI citation:

• Google reviews: going from no reviews to a solid base of recent reviews has an outsized impact. Set up a review request sequence for your existing customers. AI systems cross-reference review platforms when deciding whether to cite a business.
• Directory profiles: complete, accurate listings on Golden Pages, Yelp, Clutch (for professional services), TripAdvisor (for hospitality), LinkedIn, and any sector-specific directories (Law Society, Fáilte Ireland, RECI). AI systems cite these directories directly.
• Press and earned media: a mention in your local newspaper, a feature in an industry publication, or a guest article on a relevant website. Citation studies consistently find earned coverage like this cited far more often than brands' own websites, making it one of the strongest signals available to local businesses.
• Google Business Profile first, then Bing Places and Apple Business Connect: your Business Profile receives 28.6% of local AI citations (BrightLocal, September 2026). Bing Places feeds Copilot and Apple Business Connect feeds Siri; they are cheap to claim, though unmeasured.
• LinkedIn, YouTube and community threads: LinkedIn appears in 14.3% of ChatGPT search answers (Semrush, March 2026), and YouTube mentions are the strongest single correlate of AI visibility measured. Reddit threads are a strong source for Google AI Mode and Perplexity, though not for ChatGPT since August 2026.

AI Search Optimisation checklist for local businesses

Use this checklist to assess where your site currently stands. It covers all three pillars (SEO foundation, AI citation readiness, off-site authority):

• robots.txt allows the answer crawlers (OAI-SearchBot, ChatGPT-User, Claude-SearchBot, PerplexityBot, Googlebot, Bingbot), and your CDN or firewall lets them through
• XML sitemap exists, referenced in robots.txt, and submitted to both Google and Bing
• Google Business Profile claimed, verified, and fully completed
• Bing Places for Business claimed and matching your Business Profile exactly (for Copilot)
• Bing Webmaster Tools verified and its AI Performance report checked, and the Generative AI report in Search Console recorded
• IndexNow configured (hygiene: it speeds up Bing and Copilot, not Google or ChatGPT)
• Listed on at least 6 directories relevant to your sector
• Every key page says what you are and answers its question in the first 100 words
• Headings on service pages match real customer questions
• Every service page has a visible question-and-answer section with at least five questions
• At least one comparison table on service or product pages
• LocalBusiness schema accurate and matching the visible page (hygiene)
• About page clearly identifies your business, team, and credentials
• Honest published and "last updated" dates on content pages, backed by real updates
• At least 5 Google reviews, with a steady flow of new ones
• At least 5 answer content pages targeting specific AI queries for your sector
• At least one press mention, or a place on a "best of" roundup that AI assistants cite