A comparison page that AI answers cite opens with a one-line verdict, then compares the tools in a table of dated facts, each linked to its source, and says plainly who each option suits. Name competitors fairly, include where they win, and update the page whenever prices or features change.
Disclosure: AI SEO Gazette runs a tools directory with free listings and a paid $15 Featured listing. Where this guide mentions it, the same advice applies to any directory you choose.
Key takeaways
- In a Ten Speed study of 170 buyer-stage prompts (September 2026), comparison prompts were 20% of the set but drew close to 27% of AI citations.
- Product pages earned 24.1% of citations in the same study and comparison pages about 13%. Buyers comparing tools see these pages quoted.
- Google’s review guidance asks for evidence, measurements, pros and cons, and comparisons with other options.
- The FTC encourages naming competitors in comparisons, as long as claims are truthful and the basis of comparison is clear.
- A fair page beats a sales page. If the competitor wins on a point, say so.
Why do comparison pages get cited in AI answers?
When a buyer asks ChatGPT or Perplexity “X vs Y for small teams”, the assistant needs a page that already compares the two. In the Ten Speed study, written up by Search Engine Journal after six fact-check questions, comparison-format prompts produced about 1.33 times their share of citations. The study tracked 7,387 citation appearances across ChatGPT, Perplexity, Claude and Gemini using Peec AI. Its authors note the data was reported descriptively, so read it as a strong signal rather than proof.
A separate Discovered Labs analysis of about 2 million AI citations (last updated August 2026) found the single strongest factor was how closely a page’s content matched the prompt. Comparison pages did not show extra lift on their own in that model, which is a useful warning: a “vs” page helps only when it answers the exact comparison the buyer asks.
What should a comparison page include?
| Section | What to write |
|---|---|
| Verdict | One or two sentences: who should pick which tool |
| Comparison table | Price, free plan, main limits, key features, integrations, support. Every fact dated and linked |
| Where each tool wins | Short sections for your tool and for each competitor, with honest strengths |
| Who it suits | Use cases or team types for each option |
| How you tested | What you checked, when, and on which plan |
| Switching notes | Import, export and migration steps, if you are the alternative |
| FAQ | The follow-up questions buyers ask, such as pricing changes or contract terms |
How do you keep a comparison page fair?
Google’s guidance on high quality reviews sets a useful bar for any comparison: show evidence of your own experience, share quantitative measurements, explain what sets each product apart, cover comparable options and weigh pros and cons. The FTC statement on comparative advertising encourages naming competitors, provided comparisons are truthful, not deceptive, and the basis of comparison is clear. Four rules follow:
- Link every competitor fact to the competitor’s own page, and add the date you checked it.
- Compare like with like: the same plan level, the same billing period, the same limits.
- Include at least one point where the competitor wins. Readers and AI systems trust a page more when it is not one-sided.
- Disclose your interest. If you make one of the tools, say so at the top.
Worked example: a comparison table with dated, linked facts
Here is a small comparison built the right way, using a topic many tool owners ask about: how ChatGPT search and Perplexity crawl websites. Every cell comes from each company’s own documentation, checked on 21 November 2026.
| Question | ChatGPT search (OpenAI docs) | Perplexity (Perplexity docs) |
|---|---|---|
| Crawler for search results | OAI-SearchBot | PerplexityBot |
| Separate training crawler | GPTBot | None declared. PerplexityBot “is not used to crawl content for AI foundation models” |
| Bot for user-requested visits | ChatGPT-User. “robots.txt rules may not apply” | Perplexity-User. Generally ignores robots.txt |
| Published IP ranges | Yes | Yes, for both bots |
| Time for robots.txt changes | About 24 hours | Up to 24 hours |
| Effect of blocking the search crawler | Not shown in ChatGPT search answers, can still appear as navigational links | Cannot appear in Perplexity search results |
Verdict: the two work the same way for site owners: allow the search crawler to be cited, and decide on training separately. The difference is that OpenAI runs a declared training crawler and Perplexity says it does not.
Notice what makes this table citable. Each row answers one question a reader could ask. Each fact can be checked in one click. The verdict follows from the table instead of from opinion. Build your own “X vs Y” page the same way, swapping in price, plans and features.
Comparison page template
H1: [Your tool] vs [Competitor]: [main difference in a few words]
Verdict (2 sentences): who should choose each.
Disclosure: we make [Your tool].
Table: rows = buyer questions, columns = tools, each cell dated and linked.
Where [Competitor] wins: 2 to 3 honest points.
Where [Your tool] wins: 2 to 3 points with proof.
Who each tool suits: 3 to 4 use cases.
How we compared: plans, dates, what we tested.
Switching from [Competitor]: import steps.
FAQ: 3 to 5 follow-up questions.
Last checked: [date].Comparison or alternatives page?
Use a “X vs Y” page when buyers already have two names and want a decision. Use an “X alternatives” page when buyers are unhappy with a known tool and want options. An alternatives page lists several tools, including yours, with a short fair summary of each. Both work best on your own site and as guest content on sites your buyers read.
How often should you update a comparison page?
Every time a price or plan changes, and at least once a quarter. Show the date you last checked. Google’s helpful content guidance warns against changing dates without changing content, so only update the date when you have actually rechecked the facts.
Where AI gets it wrong
AI writing tools are risky for comparison pages because they fill gaps with plausible guesses. Ask one to compare two tools and it will often state a competitor’s price or feature limit with no source, and those details are frequently out of date. Treat any AI draft as a list of questions, then fill every cell yourself from the competitor’s own pricing and docs pages.
Frequently asked questions
Is it legal to name competitors on my site?
In the US, the FTC encourages naming competitors in comparisons as long as claims are truthful, not deceptive and clearly based. Laws differ by country, so check local rules for trademark use.
Should I put my own tool first?
Yes, if you disclose that you make it. Readers expect it. What matters is that the facts about the competitor are fair and sourced.
What if a competitor asks me to change the page?
Check their claim against their own current pages. If you were wrong, fix it and update the date. If you were right, keep the source link.
How many tools should an alternatives page cover?
Enough to give a real choice, usually five to ten. Cover each one honestly rather than listing dozens with one line each.
Sources
- Search Engine Journal (September 2026), Product pages get 24% of AI citations, Reddit and YouTube 4% (Ten Speed study)
- Discovered Labs (August 2026), What actually drives AI citations
- Google Search Central, Write high quality reviews
- US Federal Trade Commission, Statement of policy regarding comparative advertising
- OpenAI, Overview of OpenAI crawlers
- Perplexity, Perplexity crawlers
Related: AI brand mentions and how to get recommended by ChatGPT.
Part of our SaaS SEO guide for AI tools. More guides in Tool Owner Resources.
