How to Get Your Business Recommended by ChatGPT in 2026 (Complete AEO Playbook)
TL;DR: ChatGPT, Perplexity, Claude, and Gemini decide which businesses to recommend using three signals: (1) structured data on your website, (2) consistent mentions across third-party sites, and (3) explicit permissions for AI crawlers. Most small businesses fail on all three. This guide walks through exactly what to add to your site—llms.txt, schema markup, AI crawler permissions, and citation building—so AI assistants cite you when someone asks “who should I hire near me?”
Why This Matters in 2026
In 2024, “best plumber near me” sent users to Google, who sent them to ten blue links. In 2026, the same query gets answered directly by ChatGPT, Perplexity, or Google AI Overviews—with three recommendations and a summary. Users click through to one of the three, or they call the business the AI named first.
If your business isn’t in that shortlist, you don’t exist.
This is called Answer Engine Optimization (AEO), and it’s a separate discipline from traditional SEO. Google ranks pages. Answer engines recommend entities. The mechanics are different.
How ChatGPT Decides Which Businesses to Recommend
When a user asks ChatGPT “find me a good dentist in Austin,” the model does roughly this:
- Web search — ChatGPT queries Bing (its default search provider) for relevant results.
- Entity extraction — it reads the top pages and extracts named businesses, pulling structured data (schema.org) where available.
- Cross-reference — it checks how often each business is mentioned across independent sources: review sites, listicles, directories, Reddit, local news.
- Rank by confidence — businesses with consistent signals (same name, address, phone, services) across multiple sources get recommended. Businesses with thin or inconsistent data get skipped.
Perplexity, Gemini, and Claude work similarly. The implication: your website is necessary but not sufficient. You also need third-party mentions and clean, machine-readable data.
The 7 Things Your Business Website Needs
1. A /llms.txt file
llms.txt is a plain-text summary of your business written for AI systems. Place it at yoursite.com/llms.txt. Include:
- One-line business description
- Services you offer
- Location and hours
- Contact info
- Pricing (if applicable)
- Key FAQs
Most websites don’t have this file. Adding one is a low-effort, high-leverage signal that puts you ahead of 99% of local businesses.
See our template below or use our free AEO Score tool to check whether your site has one.
2. LocalBusiness schema in JSON-LD
Schema.org structured data tells AI exactly what your business is, where it is, and what it does. Add this to your site’s <head>:
<script type="application/ld+json">
{
"@context": "https://schema.org",
"@type": "LocalBusiness",
"name": "Your Business Name",
"image": "https://yoursite.com/logo.jpg",
"address": {
"@type": "PostalAddress",
"streetAddress": "123 Main St",
"addressLocality": "Austin",
"addressRegion": "TX",
"postalCode": "78701"
},
"telephone": "+1-512-555-0100",
"url": "https://yoursite.com",
"openingHours": "Mo-Fr 09:00-17:00"
}
</script>
If you run a specific business type (Restaurant, Dentist, Plumber), use the more specific schema—it carries more signal.
3. FAQPage schema with real questions
AI assistants heavily weight FAQ schema because it’s structured Q&A—easy to extract and cite. Add 5–10 Q&A pairs covering questions customers actually ask. Example structure:
{
"@type": "FAQPage",
"mainEntity": [{
"@type": "Question",
"name": "Do you offer emergency service?",
"acceptedAnswer": {
"@type": "Answer",
"text": "Yes, we offer 24/7 emergency plumbing service in Austin..."
}
}]
}
Write the questions the way a customer would type them into ChatGPT, not the way your marketing team would phrase them.
4. Explicit AI crawler permissions in robots.txt
AI crawlers check robots.txt before indexing. Many sites inadvertently block them. Add:
User-agent: GPTBot
Allow: /
User-agent: ClaudeBot
Allow: /
User-agent: PerplexityBot
Allow: /
User-agent: Google-Extended
Allow: /
Google-Extended controls whether Google uses your content in AI Overviews. Many SEO guides tell you to block it—don’t, unless you have a strong reason. Blocking it removes you from Google’s AI answers.
5. Consistent NAP (Name, Address, Phone) across the web
AI cross-references your business across sources. If your address is “123 Main Street” on your site and “123 Main St.” on Yelp and “123 Main” on Google, the model may treat these as three different businesses and trust none of them.
Pick one canonical format. Use it everywhere: Google Business Profile, Yelp, Facebook, Apple Maps, Bing Places, industry directories.
6. Citations on third-party sites
This is the single hardest and highest-leverage step. AI models trust mentions from other sites more than anything on your own site. Priorities:
- Google Business Profile — still the highest-authority local signal
- Industry directories — Angi, HomeAdvisor, Yelp, Tripadvisor, Avvo (depending on industry)
- Reddit — ChatGPT cites Reddit extensively. Authentic participation in your niche subreddit pays off over months
- Local news coverage — a single local news mention is worth 50 directory listings
- Listicles — “best dentists in Austin” articles. Pitch bloggers. Offer data, quotes, or a useful angle
You can’t fake this. AI models detect coordinated link-building and discount it.
7. A SpeakableSpecification for voice assistants
If you want Siri, Alexa, or Google Assistant to read your site aloud, mark which sections are speakable:
<script type="application/ld+json">
{
"@context": "https://schema.org",
"@type": "WebPage",
"speakable": {
"@type": "SpeakableSpecification",
"cssSelector": ["h1", ".business-summary"]
}
}
</script>
This matters less than schema and crawler permissions but takes two minutes to add.
llms.txt Template
Copy this, replace the brackets, and upload to yoursite.com/llms.txt:
# [Business Name]
> [One-sentence description: what you do, where, for whom]
## Services
- [Service 1]
- [Service 2]
- [Service 3]
## Location & Hours
- Address: [Full address]
- Phone: [Phone]
- Hours: [Days and times]
## Pricing
[Starting prices or pricing approach]
## Frequently Asked Questions
**[Question a customer would ask]?**
[Short, direct answer]
**[Another question]?**
[Answer]
## Contact
- Website: [URL]
- Email: [Email]
- Phone: [Phone]
How to Check Whether You’re Discoverable
Three tests, in order of usefulness:
- Ask ChatGPT directly: “Who are the best [your service] in [your city]?” Run this three times in different sessions. If you’re not in any of the answers, you have work to do.
- Check Perplexity: Same query. Perplexity shows its sources, which tells you which pages got cited and why.
- Run an AEO audit: Our free AEO Score tool scores your site across six dimensions—llms.txt, schema, crawler permissions, NAP consistency, third-party citations, and content structure—in about 30 seconds.
The Uncomfortable Truth About Speed
On-site fixes (llms.txt, schema, robots.txt) take a few hours and produce results in 2–6 weeks as crawlers re-index. Off-site citation building (directories, Reddit, listicles) takes months and produces results in 3–12 months.
If you need AI recommendations this quarter, start with on-site fixes today. If you need them sustainably, start the citation work immediately alongside—it has the longer lead time.
Frequently Asked Questions
How long does it take to appear in ChatGPT recommendations? On-site changes show up in 2–6 weeks as AI crawlers re-index. Building enough third-party citations to get cited consistently takes 3–6 months for most local businesses.
Is AEO different from SEO? Yes. SEO optimizes for Google’s ranking algorithm and ten blue links. AEO optimizes for AI answer engines that extract entities and cite sources. Overlap exists—good SEO hygiene helps AEO—but AEO requires structured data, llms.txt, and third-party citations that traditional SEO doesn’t emphasize.
Do I need a blog to get recommended by ChatGPT? No, but useful content helps. ChatGPT cites pages that directly answer user questions. A short FAQ page with structured Q&A outperforms a long, unstructured blog post.
Will blocking AI crawlers hurt my business? Almost certainly yes, in 2026. The tradeoff made sense in 2023 when AI companies were scraping content to train models. In 2026, the same crawlers route real user traffic to your site. Blocking GPTBot means ChatGPT can’t recommend you.
Does WebZum handle all this automatically?
Yes. Every WebZum website ships with /llms.txt, /llms-full.txt, FAQ schema, LocalBusiness schema, AI crawler permissions, and SpeakableSpecification built in. No configuration needed. Build your site in 5 minutes →
What to Do This Week
Ordered by leverage:
- Add
/llms.txtand LocalBusiness + FAQPage schema to your site. - Check
robots.txtto confirm GPTBot, ClaudeBot, PerplexityBot, and Google-Extended are allowed. - Standardize your NAP across Google Business Profile, Yelp, Facebook, and Apple Maps.
- Run the AEO Score audit and fix anything it flags.
- Start the citation work: pick one directory, one subreddit, and one listicle target per week.
If this sounds like a lot, it’s because it is. If you want the on-site portion handled automatically, WebZum generates all of it in 5 minutes—llms.txt, schema, crawler permissions, speakable markup, sitemap, the whole stack.
Either way, start now. The businesses that got into Google’s top 10 in 2010 had a multi-year head start on everyone who started later. The same dynamic is playing out with AI answer engines right now.