How to Get ChatGPT to Recommend Your Product
A growing share of buying decisions start with an AI assistant, not a Google search. If ChatGPT, Claude, and Perplexity don't know your product exists — or describe it incorrectly — you're invisible to that entire channel. Here's how to fix that.
Why this matters now
of informational queries now go to AI assistants instead of search engines (and growing)
organic AI traffic if your site is blocked to AI crawlers — a common misconfiguration
of startup sites are missing at least one critical AI visibility signal
7 steps to get recommended by AI assistants
Check your current AI Exposure Score
Before fixing anything, measure where you stand. Run your URL through the free AI Exposure Score audit — it shows exactly which of the 25 signals are failing and by how much. Most products score between 20–50 on first scan. The breakdown tells you exactly where points are being left on the table.
- Run a free scan at aiexposuretool.com
- Note your score across all 6 categories
- Prioritize fixes with the biggest point impact first
Make your site AI-crawlable
Before AI assistants can recommend you, they need to be able to read your site. Many products accidentally block AI crawlers in their robots.txt — either by disabling all bots, or by specifically blocking GPTBot, Claude-Web, or PerplexityBot.
- Check robots.txt and remove disallow rules for GPTBot, Claude-Web, PerplexityBot, OAI-SearchBot, anthropic-ai
- Ensure your main content pages are publicly accessible (no login walls for marketing pages)
- Verify that key landing pages return 200 status (no soft 404s or redirects loops)
Create an llms.txt file
llms.txt is a plain-text brief file at yoursite.com/llms.txt that tells AI assistants exactly what your product does, who it's for, what it costs, and where to send users. It's the single highest-ROI AI visibility action you can take — a 30-minute investment that puts your product description directly in AI context windows.
- Create /llms.txt following the llmstxt.org specification
- Include: product name as H1, one-line description as blockquote, Overview, Target Audience, Key Features, Pricing, Key URLs
- Keep it under 1,000 words — AI context windows are finite
- Optionally create /llms-full.txt for richer detail
Add JSON-LD structured data
JSON-LD schema.org markup gives AI assistants a machine-readable summary of your product — name, description, pricing, category, URL. This is the signal that most reliably causes AI assistants to describe your product correctly when they do know about it.
- Add SoftwareApplication or Product schema on your homepage
- Include: name, description, url, applicationCategory, offers (price, priceCurrency)
- Add FAQPage schema to any page with Q&A content
- Add BreadcrumbList schema to interior pages
Write content that answers comparison questions
AI assistants give recommendations when users ask comparison questions: 'What's the best tool for X?', 'How do I do Y?', 'What should I use for Z?' Your content needs to appear in the training data or current crawl results for those questions. The highest-ROI content type is a comparison page in your own category — not naming competitors, but defining your category and explaining what to look for.
- Create a 'best [category] tools' or 'how to choose a [category] tool' page on your own site
- Write about your category from first principles — not just your product
- Include your product's differentiating criteria prominently
- Target long-tail comparison queries: 'AI-built app security check', 'answer engine optimization tool'
Build trust signals AI systems recognize
AI systems weigh trust signals heavily when deciding what to recommend — especially for commercial products. Backlinks from authoritative domains, structured author/publisher information, and consistent brand mentions across the web all raise confidence that your product is real and recommendable.
- Get coverage on Product Hunt, Hacker News, indie directories, and startup blogs
- Add Organization schema on your homepage (name, url, description, sameAs links to social)
- Ensure consistent brand name across domain, social handles, and schema markup
- Add a factual, crawlable About page with clear publisher identity
Verify and monitor your AI visibility
AI visibility changes as models get updated, new crawls happen, and competitors create competing content. Check your AI Exposure Score monthly — a score drop often signals a crawl issue, a blocked bot, or a competitor moving into your category.
- Re-run your AI Exposure Score scan after making fixes
- Manually ask ChatGPT, Claude, and Perplexity about your product category monthly
- Monitor whether AI assistants describe your product correctly
- Track inbound AI-referred traffic in your analytics
Key signals by category
These are the four signal groups that have the highest impact on whether AI assistants recommend your product. Check all of them.
AI Crawlability
- robots.txt allows AI bots
- llms.txt present
- llms-full.txt present
- No login walls on marketing pages
Structured Data
- SoftwareApplication or Product schema
- Organization schema
- FAQPage schema where applicable
- BreadcrumbList on interior pages
Content Signals
- Clear product description on homepage
- Comparison/category content
- Pricing publicly accessible
- Target audience explicitly stated
Trust & Authority
- External backlinks from relevant domains
- Product Hunt / HN listing
- Consistent brand name across web
- Author/publisher identity in schema
Questions
How long does it take for ChatGPT to start recommending my product?
There's no fixed timeline — it depends on when OpenAI's crawlers next visit your site and when models are updated. Implementing llms.txt and JSON-LD schema can show results in 2–8 weeks for products with crawlable sites. Products that are already being crawled but described incorrectly can see improvements faster once the correct structured data is in place.
Does ranking in Google affect whether AI recommends me?
Not directly. Google PageRank and AI citation are different systems. AI assistants like Perplexity use real-time web search, so search ranking matters there. But ChatGPT and Claude primarily draw on training data and direct crawl — a product with low Google rankings but strong llms.txt, JSON-LD, and authoritative inbound links can still get recommended.
What's the difference between AI SEO and traditional SEO?
Traditional SEO optimizes for keyword-ranked blue links. AI SEO (AEO — Answer Engine Optimization) optimizes for AI assistants that synthesize answers and recommend specific products. The signals overlap (authoritative content, backlinks) but AEO adds AI-specific signals: llms.txt, JSON-LD schema for software products, AI crawler access, and structured comparison content. See our comparison guide for a full breakdown.
Is getting a high AI Exposure Score enough?
The score measures signals your site currently has — it correlates with AI visibility but isn't the only factor. Training data recency, brand mention volume, and category competition all affect whether AI systems choose to recommend you. A high score is a necessary condition, not a sufficient one. The score helps you eliminate the preventable barriers first.
See where your product stands right now
Free AI Exposure Score scan — 25 signals, 30 seconds. Find out exactly what's stopping AI assistants from recommending you.
Run free AI Exposure Score