Who wins the AI Overview: 4 SaaS categories, measured
The brand that AI engines cite most in your category is probably not the one with the biggest market share — it's the one that publishes the clearest, most-linkable answers.
The finding
In 3 of 4 categories, the most-cited brand in AI answers ranked outside the top 2 by market share. Answer visibility is winnable independent of size.
Share of AI answers vs. market position, by category
| Category | Most-cited brand | Its share of AI answers | Market-share rank | Gap |
|---|---|---|---|---|
| Project management | Height | 34% | #5 | +4 spots |
| Email marketing | Beehiiv | 29% | #3 | +2 spots |
| Observability | Grafana | 41% | #2 | +1 spot |
| Customer support | Plain | 26% | #6 | +5 spots |
Based on n=480 prompts (120 per category) across ChatGPT, Perplexity, Gemini, and Google AI Overviews, run 2026-08-04 to 2026-08-08.
Why it's happening
AI answer engines don't rank brands by revenue — they assemble answers from the sources they can parse, trust, and quote. Three things move the needle:
- Extractable structure. Pages that state a claim, then back it with a table or a number, get lifted verbatim. Marketing pages full of adjectives don't.
- Third-party corroboration. Being named in comparison posts, docs, and forum answers matters more than your own homepage copy.
- Freshness with a date. Engines favor sources that show when a fact was true. Undated pages get discounted.
The smaller brands winning here aren't outspending anyone. They're publishing answer-shaped content that's trivially easy to cite.
What to do this week
- Pick your 10 highest-intent category prompts (e.g. "best X for Y") and check, today, whether you appear in each engine's answer.
- On your top 3 comparison pages, put the claim + a number/table in the first screen — before any narrative.
- Add a visible
Last updateddate and a one-paragraph methodology to any page that states data. - Earn one new third-party citation this week (a docs mention, a comparison post, a credible forum answer).
Also worth watching
Perplexity now shows citation counts on hover
A quiet UI change worth noting: Perplexity surfaces how many sources back each claim. Answers assembled from more corroborating sources appear to rank higher — another point for third-party citations over self-published copy.
Get your brand's AI-visibility report Free. Your domain is the only input. Delivered in 24–48 hours.
The full picture
Market-share leaders assume their size carries into AI answers. It doesn’t. When an engine builds an answer to “best project management tool for agencies,” it isn’t consulting a market-share table — it’s stitching together the most quotable sources it can find on the open web.
That’s why Height — a distant #5 by market share — shows up in a third of project-management answers. Their comparison pages lead with a claim and a table, they timestamp their data, and they’re referenced across enough third-party posts that engines treat them as a safe source to quote.
The uncomfortable, hopeful implication for everyone else: AI answer visibility is not pay-to-win, and it’s not size-to-win. It’s structure-to-win. The brands that publish the clearest answers get cited — regardless of how big they are.
Next week: we break down the exact page structures that got lifted verbatim into AI answers, with before/after examples.
Methodology
For each category we drafted 120 buyer-intent prompts spanning discovery ("best tools for…"), comparison ("X vs Y"), and validation ("is X good for…"). Each prompt was run once per engine (ChatGPT, Perplexity, Gemini, Google AI Overviews) from a clean, unauthenticated US session between 2026-08-04 and 2026-08-08. A brand counted as "cited" if it was named in the answer text or appeared as a linked source. Share of AI answers is the percentage of a category's prompts in which a brand was cited, averaged across the four engines. Market-share ranks are from public 2025 category reports. This is a point-in-time snapshot; engines change, and so will these numbers.
Published August 12, 2026. Data by Morning Crawl.
FAQ
What counts as being 'cited' in an AI answer?
A brand is counted as cited when it is named in the generated answer text or appears as a linked source. Both self-published and third-party sources count.
Which engines did you measure?
ChatGPT, Perplexity, Google Gemini, and Google AI Overviews — the four engines that most often surface for buyer-intent queries.
How often is this data refreshed?
Morning Crawl publishes a new measured edition every week. Each edition is a dated, point-in-time snapshot.