AI Visibility Benchmark: How Often Does AI Recommend SaaS Brands? (2026 Data)
We asked an AI answer engine about 101 SaaS brands across 23 categories — the brand itself, its reviews, and its alternatives — and recorded which sources it actually cited. This is the anonymous, aggregate picture of AI search visibility in mid-2026.
Data collected July 2026 · 101 brands · 2,469 structured citations · engine: Perplexity (sonar)
The Most Cited AI Sources comes from this same measurement — one run over these 101 brands, read two ways. Two pages, not two studies: reading both is not corroboration.
of brands get cited when you ask AI about them by name
get cited when buyers ask AI for alternatives in their category
of brands AI clearly knows still vanish from alternatives answers
of everything AI cites is third-party content, not the brand's own site
The visibility slope: AI knows you — until money is on the line
Ask AI about a brand by name and it cites that brand's own site 91% of the time. Ask for the brand's reviews and the citation rate drops to 56%. Ask for its alternatives — the question a buyer asks right before choosing — and it drops to 55%. The closer a query gets to purchase intent, the less control a brand has over the answer.
The alternatives gap: 42% of known brands vanish
39 of 92 name-recognized brands were not cited on the alternatives query for their own category
We looked only at the brands the AI engine clearly knows — the ones it cites when asked about them by name. When we then asked for alternatives in their category, 42% of those same brands were nowhere in the cited sources.
Being well-known is not enough. AI assembles alternatives answers from third-party comparison content — community threads, listicles, review platforms — and a brand that is absent from those surfaces is absent from the answer, no matter how big its name is. That is a property of the mechanism, not a scoreboard: it hits famous and niche brands alike.
How many queries does a brand actually win?
Out of the three queries per brand, 42 of 101 brands were cited on all three — and 5 were cited on none. The average visibility score is 68 out of 100.
Coverage is wildly uneven across categories
Average visibility score — the share of the three queries on which a brand gets cited — ranges from 33 to 93 across categories. Categories with strong community and review footprints sit near full coverage; categories where buying research happens elsewhere lag far behind.
| Category | Brands sampled | Avg visibility score | Cited on alternatives |
|---|---|---|---|
| CRM | 9 | 52 | 56% |
| HR | 9 | 71 | 56% |
| Project management | 9 | 63 | 67% |
| Design | 6 | 56 | 50% |
| Productivity | 6 | 55 | 33% |
| Analytics | 5 | 93 | 80% |
| Finance | 5 | 40 | 20% |
| Knowledge base | 5 | 87 | 80% |
| Monitoring | 5 | 87 | 80% |
| Video | 5 | 33 | 0% |
| Automation | 4 | 42 | 25% |
| Dev platform | 4 | 75 | 50% |
| Forms | 4 | 92 | 75% |
Showing 13 of 23 categories — those with at least 4 brands sampled. Scores are category averages; no per-brand data is published.
Domain Rating does not predict AI citation
Every brand in this run has an Ahrefs Domain Rating, so the two can be crossed directly. Sorted into four bands by DR, the share of prompts each band was cited on barely moves — and the highest-DR band sits at 0.600, below the lowest-DR band's 0.641. This is the premise the product rests on, so the three limits below carry the same weight as the finding.
| DR band | Brands | Mean citation rate |
|---|---|---|
| DR 59–76 | 26 | 0.641 |
| DR 77–81 | 25 | 0.760 |
| DR 82–89 | 30 | 0.689 |
| DR 90–95 | 20 | 0.600 |
Bands are cut on the DR value, not on rank: each band takes every brand above the previous band's top DR up to and including its own. That keeps brands with identical DR in the same band — 10 of these 101 share DR 81 — at the cost of unequal band sizes. The choice matters: splitting into four equal-sized groups by rank instead moves the top band from 0.600 to 0.641, so the cut points are published rather than implied.
What this does and does not show
- The sample runs from DR 59 to DR 95 — every brand in it is already a high-authority domain. So the finding is that among domains that are already strong, DR does not predict AI citation. It is not evidence that DR and AI visibility are unrelated; this run contains no weak domains to say that with.
- Each brand was asked 3 prompts, so its citation rate can only take 4 values: 0.0000 / 0.3333 / 0.6667 / 1.0000. A measure that coarse attenuates any correlation on its own.
- At n=101, this sample can detect a correlation of |r|=0.276 with 80% power. It therefore rules out a moderate or strong relationship, and does not rule out a weak one — the 95% interval reaches 0.220.
Pearson r = 0.025 across all 101 brands, 95% CI [-0.171, 0.220]. Citation rates come from the July 2026 run; Domain Ratings are the snapshot of August 5, 2026. The two were measured a month apart and are dated separately here for that reason.
Where AI actually looks: Reddit and YouTube dominate
Across 2,469 structured citations, the single biggest sources are Reddit (9.1%) and YouTube (8%) — ahead of any review platform. And only 10% of all citations point at brand-owned domains: about 90% of what AI reads about a brand is content the brand does not control.
| Source | Share of all citations | Cited for # of brands |
|---|---|---|
| reddit.com | 9.1% | 98 |
| youtube.com | 8% | 89 |
| g2.com | 3.9% | 73 |
| trustpilot.com | 3.7% | 73 |
| en.wikipedia.org | 2.5% | 55 |
Top sources on reviews queries
- 1reddit.com11.4%
- 2youtube.com10.7%
- 3trustpilot.com8.7%
- 4g2.com7.5%
- 5glassdoor.com4.2%
Top sources on alternatives queries
- 1reddit.com11%
- 2youtube.com4.8%
- 3g2.com2.3%
- 4thedigitalprojectmanager.com1.4%
- 5facebook.com1%
Only hosts cited across at least 5 different brands are listed, so no single brand's footprint can be reverse-engineered.
What this means if you run a SaaS
Optimize for the buying queries, not the brand query
Brand-name visibility is nearly free (91%). The reviews and alternatives queries are where roughly half of brands disappear — and where buyers actually decide.
Win the third-party surfaces AI cites
Reddit threads, YouTube reviews, comparison listicles and review platforms make up most of what AI reads. Around 90% of citations are content you don't own — earn presence there instead of only polishing your own site.
Measure per query type, not overall
A brand can be fully visible on its name and invisible on alternatives. Track brand, reviews and alternatives queries separately — averages hide exactly the gap that costs you buyers.
Methodology
Sample: 101 established SaaS brands across 23 categories (project management, CRM, HR, analytics, design, finance and more). SEO/GEO tools were excluded from the seed list to avoid self-interest. Data collected July 2026.
For each brand we asked Perplexity (sonar) three questions a real buyer asks — the brand name, "[brand] reviews", and "[brand] alternatives" — and recorded the engine's structured citations. A brand counts as cited on a query only when its own domain appears among those structured sources; the answer prose is never parsed.
Honesty rules: failed calls are recorded as unknown and excluded from every denominator — never counted as "not cited". Rerun rows are deduplicated keeping the latest attempt. In this run, all 101 brands completed all three queries.
Anonymity: only aggregates are published. Categories appear only with 4+ brands sampled; source hosts only when cited across 5+ different brands. Every number on this page is reproducible from the raw run by the committed aggregation script.
Which AI crawlers actually fetch pages — a separate measurement, on our own server →
FAQ
What is an AI visibility benchmark?+
An AI visibility benchmark measures how often AI answer engines cite or recommend brands when users ask about them. This benchmark covers 101 SaaS brands across 23 categories, measured on Perplexity (sonar) in July 2026 with three buyer-style queries per brand: the brand name, its reviews, and its alternatives.
How is AI visibility measured here?+
A brand counts as visible on a query when its own domain appears in the engine's structured citations for that query. In this dataset, 91% of brands are cited on their brand-name query, 56% on reviews queries and 55% on alternatives queries. Failed calls are excluded from denominators, never counted as misses.
Why doesn't this benchmark name or rank the brands?+
By design. The purpose is to show how AI recommendation behaves as a mechanism — citation rates, the alternatives gap, category differences and dominant sources — not to score individual companies. Only aggregate, anonymous numbers are published here, and source hosts are shown only when they appear across many different brands. We never share one brand's result with anyone else — but if it is your brand, you can ask us for your own row.
Can I check my own brand's AI visibility?+
Yes. TriRank's free AI visibility audit runs the same kind of structured-citation check on your own domain: whether AI cites you, which sources it reads instead, and how you compare to competitors. The audit is free.
Check your own AI visibility — free
Run the same structured-citation check on your domain: see whether AI cites you on brand, reviews and alternatives queries, and where it looks instead.