This isn't a theory post. No "AI is changing everything" preambles. You're here because your brand doesn't show up when people ask ChatGPT for recommendations, and you want to fix that.
Here are eight specific tactics, ordered by impact. Each one includes exactly what to do, not just what to think about.
Tactic 1: Create Ungated Comparison Content
Why it works: When someone asks ChatGPT "What's the best email marketing tool?", the model synthesizes information from across the web. Comparison articles are the richest source of "Brand X is good for Y" signals that AI platforms use to form recommendations.
Exactly what to do:
Create dedicated pages on your site for each major comparison:
yoursite.com/compare/your-brand-vs-competitor-ayoursite.com/blog/best-[category]-tools-2026yoursite.com/compare/[category]-comparison
Template for comparison pages:
## [Your Brand] vs [Competitor]
### Quick Summary
[Your Brand] is best for [use case]. [Competitor] is best for [different use case].
### Feature Comparison
[Detailed, honest table comparing specific features]
### Pricing
[Transparent pricing comparison]
### Who Should Choose [Your Brand]
[Specific audience and use cases]
### Who Should Choose [Competitor]
[Be honest — this builds authority]
The key: be genuinely fair. AI platforms can detect one-sided comparisons (or rather, balanced content from third parties will outweigh biased content). Acknowledge competitor strengths. This counterintuitively helps you — AI treats balanced content as more authoritative.
Slack does this well. Their comparison pages against Microsoft Teams are detailed, fair, and clearly state where each tool wins. AI platforms reference these pages directly.
Tactic 2: Get Listed on Major Review and Comparison Sites
Why it works: G2, Capterra, TrustRadius, Product Hunt, and similar sites are high-authority sources that AI platforms weight heavily. When multiple trusted third-party sites list your product with positive reviews, AI forms a consensus view that your brand is credible.
Exactly what to do:
- Claim your profiles on G2, Capterra, TrustRadius, Product Hunt, AlternativeTo, and SaaSWorthy
- Complete every field — description, features, pricing, screenshots, integrations. Incomplete profiles get skipped.
- Run a review campaign: Email your happiest customers and ask for reviews. Aim for 50+ reviews on G2 — that's the threshold where you start appearing in "best of" categories.
- Respond to every review — positive and negative. Active profiles rank higher.
- Keep profiles updated — refresh features, pricing, and screenshots quarterly.
HubSpot has over 12,000 reviews on G2. You don't need that many. But the difference between 10 reviews and 100 reviews is the difference between AI knowing your product exists and AI recommending it.
Tactic 3: Build Schema Markup and JSON-LD
Why it works: Structured data gives AI platforms machine-readable facts about your product. Instead of inferring from messy HTML that you're "probably some kind of project management tool," JSON-LD tells them definitively: "This is a SoftwareApplication in the ProjectManagement category, rated 4.6/5, priced at $12/month."
Exactly what to do:
Use the JSON-LD Generator to create markup for:
Homepage — Organization + SoftwareApplication:
{
"@type": "SoftwareApplication",
"name": "Your Product",
"applicationCategory": "BusinessApplication",
"operatingSystem": "Web",
"offers": {
"@type": "Offer",
"price": "29",
"priceCurrency": "USD"
}
}
FAQ pages — FAQPage schema: Add FAQ schema to your pricing page, feature pages, and any page with a Q&A section.
Product pages — Product schema: Include ratings, reviews count, and feature descriptions.
Add the generated JSON-LD to the <head> section of each relevant page. Validate it with Google's Rich Results Test.
Tactic 4: Create an LLMs.txt File
Why it works: LLMs.txt is an emerging standard (similar to robots.txt) that gives AI crawlers a structured overview of your site specifically designed for language model consumption. It's a direct channel to tell AI what your product is and what content matters most.
Exactly what to do:
- Go to the LLMs.txt Generator
- Enter your site URL
- Review and customize the generated file
- Upload it to
yoursite.com/llms.txt
Your LLMs.txt should include:
- A clear product description
- Links to your most important pages (product, pricing, docs, comparisons)
- Your target audience and key use cases
- Feature highlights in plain language
This takes 15 minutes and immediately makes your site more legible to AI crawlers. For more detail on how this works, read how ChatGPT chooses which brands to recommend.
Tactic 5: Publish Strategic "Best X" Content
Why it works: When AI platforms look for answers to "What's the best X?", they aggregate from articles that literally answer that question. If your brand publishes authoritative "best of" content that includes your product (fairly, among competitors), you create a high-quality source that AI can reference.
Exactly what to do:
Create 3-5 definitive list articles:
- "Best [Your Category] Tools for [Specific Audience] in 2026"
- "Best [Your Category] for [Use Case]"
- "Top [Your Category] Alternatives to [Dominant Competitor]"
Critical rules:
- Include 7-10 products, not just yours
- Put your product where it honestly belongs, not always #1
- Write 200+ words per product with genuine pros and cons
- Include pricing, best-for use cases, and key features for each
- Update these articles quarterly
Notion publishes comparison content that positions itself honestly alongside Confluence, Coda, and others. The result: AI platforms treat Notion as an authority on the productivity tool space, not just a self-promoter.
Tactic 6: Contribute to Authoritative Publications
Why it works: AI platforms assess authority based on where information appears, not just what it says. A mention of your brand in TechCrunch, Forbes, or an industry-leading blog carries more weight than a mention on your own site.
Exactly what to do:
- Write guest posts for industry blogs where you can naturally mention your product in context
- Respond to journalist queries on platforms like Qwoted, Help a Reporter (HARO), or Terkel
- Contribute expert opinions to roundup articles in your industry
- Publish original research that gets cited — survey your users, analyze industry data, share benchmarks
The goal isn't to write advertorials. It's to become a cited expert in your space. When Stripe's CEO is quoted in a fintech article alongside a mention of Stripe's payment infrastructure, that signal propagates through AI training data.
Target: Get mentioned in 5-10 authoritative external sources per quarter.
Tactic 7: Make Your Product Documentation AI-Crawlable
Why it works: Product documentation is the richest source of specific, factual information about what your product does. If AI can read your docs, it can accurately describe your features, compare you to competitors, and recommend you for the right use cases.
Exactly what to do:
- Make all docs public — no login wall for reading documentation
- Use clean HTML with semantic headings (H1, H2, H3 hierarchy)
- Add a sitemap specifically for your docs subdomain
- Include a meta description on every doc page
- Structure content as questions and answers where appropriate:
- "How to set up email integration" → Clear heading, step-by-step answer
- "What integrations are available?" → Comprehensive list
- Ensure your docs cover competitive differentiators — don't just document how, document why your approach is different
Stripe's documentation is legendary — not because it's pretty, but because it's comprehensive, public, and perfectly structured. AI platforms know more about Stripe's capabilities than almost any other payment processor, largely because of documentation quality.
Tactic 8: Structure Content in Q&A Format
Why it works: AI platforms are fundamentally question-answering systems. Content structured as questions and answers maps directly to how AI generates responses. When your content literally contains the question "What's the best CRM for small business?" followed by a well-structured answer that includes your brand, you're giving AI a ready-made response.
Exactly what to do:
Add FAQ sections to your key pages:
- Homepage: 5-8 questions about your product category
- Product page: 8-10 questions about features, pricing, and comparisons
- Pricing page: 5-6 questions about plans, billing, and value
- Each feature page: 3-5 questions specific to that feature
Format that AI parses best:
## Frequently Asked Questions
### What is [Your Product]?
[Your Product] is a [category] tool that helps [audience] do [thing].
### How does [Your Product] compare to [Competitor]?
[Honest, balanced comparison with specific differentiators]
### Who is [Your Product] best for?
[Specific audience descriptions with use cases]
Add FAQPage schema markup to these sections (Tactic 3) for maximum impact.
Key takeaway: Every tactic in this list reinforces the others. Comparison content (Tactic 1) with schema markup (Tactic 3) in Q&A format (Tactic 8) linked from your LLMs.txt (Tactic 4) is exponentially more effective than any single tactic alone.
Measuring Your Progress
After implementing these tactics, track your results:
- Baseline first: Run your site through the AEO Score tool to get your starting score
- Monitor mentions: Use the AI Visibility Index to track brand mentions across platforms
- Re-test monthly: Run the same prompts and compare results
Expect to see movement on retrieval-augmented platforms (Perplexity) within 1-2 weeks. Training-based platforms (ChatGPT, Claude) take longer — typically 1-3 months for significant changes.
The Priority Stack
If you can only do three things this week:
- Create an LLMs.txt file (15 minutes) — Generator here
- Add JSON-LD to your homepage (30 minutes) — Generator here
- Publish one comparison page (2-3 hours) — your brand vs. your top competitor
That's a morning's work that puts you ahead of 90% of companies who haven't even started thinking about AI visibility. Then build from there.
The brands winning in AI recommendations right now aren't doing anything magical. They're doing these eight things consistently. Start today.
