AI visibility is how likely a page is to be read, understood, and cited by AI answer engines such as ChatGPT, Perplexity, and Google AI Overviews. The short answer is that big SaaS is weak at it: we scored the homepages of 100 leading SaaS companies with AEOCrawler's 9-dimension production scorer, and the average was 55.5 out of 100 — not one homepage reached 80. The weakest dimensions across the board were Query Coverage (32.8) and Citation Potential (33.8), and only 9.9% of homepages carry FAQ schema.
When someone asks ChatGPT, Perplexity, or Google AI Overviews "what is Notion?" or "best CRM for a small team," the page those engines most often read is the company homepage. This study measures how readable, extractable, and citable those homepages actually are — for the 100 most recognizable names in SaaS. The full ranked scoreboard is below.
The headline findings
Every finding below comes from one run of the same 9-dimension scoring model across all homepages (methodology further down).
- Not one of 91 scored homepages reached 80/100 for AI visibility. The best was Wix at 79.3.
- The average top-SaaS homepage scores 55.5/100 (median 55.6, range 32.3–79.3). This is a systemic gap, not a few laggards.
- The companies building AI are the least visible to it. Data, analytics & AI SaaS scored lowest of all ten categories (average 49.7/100) — Anthropic scored 34.4 and Hugging Face 32.3.
- The two universal weak spots are Query Coverage (avg 32.8) and Citation Potential (avg 33.8). Homepages answer few of the questions AI engines actually get asked, and give engines little they can quote.
- Only 9.9% of top SaaS homepages have FAQ schema — one of the cheapest citation levers in existence.
- 41.8% have neither Organization nor FAQ schema at all, leaving AI engines to guess at basic entity facts.
- 1 in 4 homepages (26.4%) never opens with a direct answer — the explanation of what the product is sits buried under a slogan.
- One brand — Canva — blocks AI crawlers outright in robots.txt (GPTBot, ClaudeBot, PerplexityBot, CCBot and others are disallowed sitewide). That dents its AI visibility rather than killing it: robots.txt stops model training and standards-respecting retrieval, while user-triggered fetchers may still reach the page.
The scoreboard: 91 top SaaS homepages, ranked
Each company's primary marketing homepage was scored once, on the same day, through the same scoring path. The composite is 0–100 across nine weighted dimensions; the weakest dimension shows where each homepage loses the most points. AEOCrawler builds the scoring model used here and was excluded from the ranked sample. Nine of the 100 target homepages could not be fetched from datacenter IPs and are listed after the table.
| # | Company | Category | Score | Weakest dimension |
|---|---|---|---|---|
| 1 | Wix | Design, content & web | 79.3 | Content Freshness |
| 2 | Pipedrive | CRM, sales & support | 77.6 | Conciseness |
| 3 | Auth0 | Security & identity | 76.7 | Conciseness |
| 4 | Gong | CRM, sales & support | 74.5 | Readability |
| 5 | Squarespace | Design, content & web | 74.2 | Semantic Coverage |
| 6 | Moz | Marketing & SEO | 73 | Query Coverage |
| 7 | Elementor | Design, content & web | 71.8 | Query Coverage |
| 8 | Culture Amp | HR & people | 68.2 | Citation Potential |
| 9 | Lever | HR & people | 67.8 | Readability |
| 10 | ClickUp | Productivity & collaboration | 66.9 | Query Coverage |
| 11 | Sentry | Dev tools & infrastructure | 66.8 | Query Coverage |
| 12 | HubSpot | CRM, sales & support | 66.7 | Citation Potential |
| 13 | Twilio | Dev tools & infrastructure | 66.4 | Query Coverage |
| 14 | Pinecone | Data, analytics & AI | 65.8 | Query Coverage |
| 15 | Monday.com | Productivity & collaboration | 65.2 | Query Coverage |
| 16 | Snyk | Security & identity | 64.7 | Citation Potential |
| 17 | Chargebee | Commerce & billing | 63.9 | Citation Potential |
| 18 | Amplitude | Data, analytics & AI | 63.8 | Structured Data |
| 19 | Apollo | CRM, sales & support | 63.6 | Citation Potential |
| 20 | Webflow | Design, content & web | 62.6 | Query Coverage |
| 21 | Netlify | Dev tools & infrastructure | 62.4 | Conciseness |
| 22 | Cloudflare | Dev tools & infrastructure | 62.3 | Citation Potential |
| 23 | Stripe Billing | Commerce & billing | 62 | Readability |
| 24 | Contentful | Design, content & web | 61.8 | Query Coverage |
| 25 | Zendesk | CRM, sales & support | 61.3 | Direct Answer |
| 26 | Intercom | CRM, sales & support | 61.3 | Citation Potential |
| 27 | JumpCloud | Security & identity | 61 | Citation Potential |
| 28 | Lemon Squeezy | Commerce & billing | 60.2 | Structured Data |
| 29 | 15Five | HR & people | 59.7 | Structured Data |
| 30 | Mailchimp | Marketing & SEO | 59.2 | Structured Data |
| 31 | Asana | Productivity & collaboration | 59 | Citation Potential |
| 32 | Workday | HR & people | 58.6 | Query Coverage |
| 33 | CrowdStrike | Security & identity | 58.6 | Direct Answer |
| 34 | Ghost | Design, content & web | 57.9 | Query Coverage |
| 35 | Slack | Productivity & collaboration | 57.7 | Structured Data |
| 36 | Klaviyo | Marketing & SEO | 57.4 | Direct Answer |
| 37 | Brex | Finance & operations | 57.4 | Structured Data |
| 38 | Lattice | HR & people | 57.1 | Query Coverage |
| 39 | HiBob | HR & people | 56.7 | Direct Answer |
| 40 | Recurly | Commerce & billing | 56.3 | Citation Potential |
| 41 | Mollie | Commerce & billing | 56.3 | Query Coverage |
| 42 | Front | CRM, sales & support | 56.1 | Citation Potential |
| 43 | MongoDB | Dev tools & infrastructure | 56 | Query Coverage |
| 44 | Stripe | Dev tools & infrastructure | 55.9 | Citation Potential |
| 45 | ActiveCampaign | Marketing & SEO | 55.7 | Citation Potential |
| 46 | Figma | Productivity & collaboration | 55.6 | Query Coverage |
| 47 | Greenhouse | HR & people | 55.6 | Citation Potential |
| 48 | Supabase | Dev tools & infrastructure | 55.3 | Structured Data |
| 49 | QuickBooks | Finance & operations | 54.5 | Structured Data |
| 50 | Salesforce | CRM, sales & support | 53.5 | Citation Potential |
| 51 | Buffer | Marketing & SEO | 53.4 | Citation Potential |
| 52 | Remote | HR & people | 53.2 | Query Coverage |
| 53 | Paddle | Commerce & billing | 53.1 | Query Coverage |
| 54 | dbt Labs | Data, analytics & AI | 52.9 | Structured Data |
| 55 | Databricks | Data, analytics & AI | 52.7 | Structured Data |
| 56 | Rippling | Finance & operations | 52.5 | Citation Potential |
| 57 | Xero | Finance & operations | 52.2 | Structured Data |
| 58 | Ahrefs | Marketing & SEO | 51.4 | Readability |
| 59 | Brevo | Marketing & SEO | 51.3 | Citation Potential |
| 60 | Mixpanel | Data, analytics & AI | 51.2 | Structured Data |
| 61 | Sanity | Design, content & web | 51.1 | Structured Data |
| 62 | Outreach | CRM, sales & support | 50.9 | Citation Potential |
| 63 | Semrush | Marketing & SEO | 50.4 | Citation Potential |
| 64 | Loom | Productivity & collaboration | 50.2 | Citation Potential |
| 65 | Vanta | Security & identity | 50.2 | Structured Data |
| 66 | Coda | Productivity & collaboration | 49.7 | Structured Data |
| 67 | Okta | Security & identity | 49.6 | Direct Answer |
| 68 | BigCommerce | Commerce & billing | 48.9 | Direct Answer |
| 69 | Bill.com | Finance & operations | 48.8 | Structured Data |
| 70 | Tailscale | Security & identity | 48.4 | Structured Data |
| 71 | Segment | Data, analytics & AI | 48.1 | Structured Data |
| 72 | FreshBooks | Finance & operations | 48.1 | Structured Data |
| 73 | Notion | Productivity & collaboration | 47.7 | Structured Data |
| 74 | Dashlane | Security & identity | 47.6 | Structured Data |
| 75 | Airtable | Productivity & collaboration | 47.5 | Structured Data |
| 76 | Deel | Finance & operations | 47.5 | Structured Data |
| 77 | Miro | Productivity & collaboration | 47.4 | Structured Data |
| 78 | Shopify | Commerce & billing | 46.5 | Query Coverage |
| 79 | Snowflake | Data, analytics & AI | 46.4 | Structured Data |
| 80 | Framer | Design, content & web | 46.2 | Query Coverage |
| 81 | Hootsuite | Marketing & SEO | 44.7 | Structured Data |
| 82 | Builder.io | Design, content & web | 44.2 | Structured Data |
| 83 | Ramp | Finance & operations | 43.4 | Structured Data |
| 84 | Postman | Dev tools & infrastructure | 42.2 | Query Coverage |
| 85 | Gumroad | Commerce & billing | 42.2 | Structured Data |
| 86 | Square | Commerce & billing | 42.1 | Structured Data |
| 87 | 1Password | Security & identity | 38 | Structured Data |
| 88 | Vercel | Dev tools & infrastructure | 36.4 | Citation Potential |
| 89 | Anthropic | Data, analytics & AI | 34.4 | Structured Data |
| 90 | Datadog | Dev tools & infrastructure | 32.8 | Structured Data |
| 91 | Hugging Face | Data, analytics & AI | 32.3 | Structured Data |
Not scored (9 of 100): OpenAI, Freshworks, ConvertKit, Gusto, Expensify, BambooHR, Drata, Canva, and Personio. These homepages blocked our datacenter crawler (WAF or rate limiting). Only Canva also blocks AI crawlers in robots.txt; the other eight allow them — BambooHR explicitly allows GPTBot and ClaudeBot — so their absence here says nothing about their AI visibility.
Which SaaS categories are most AI-visible
| Category | Average score |
|---|---|
| CRM, sales & support | 62.8 |
| Design, content & web | 61.0 |
| HR & people | 59.6 |
| Marketing & SEO | 55.2 |
| Security & identity | 55.0 |
| Productivity & collaboration | 54.7 |
| Dev tools & infrastructure | 53.7 |
| Commerce & billing | 53.2 |
| Finance & operations | 50.6 |
| Data, analytics & AI | 49.7 |
Two things stand out.
CRM and sales SaaS lead. These companies live off inbound comparison traffic ("best CRM for X"), and it shows: their homepages answer questions, carry schema, and open with plain-language definitions. Pipedrive (77.6) is the category's model citizen.
AI companies come last. The category that includes Anthropic, Hugging Face, Databricks, and Snowflake averaged 49.7 — the worst of all ten. The pattern is consistent: developer-brand homepages lean on product screenshots, animated demos, and slogans, and skip the structured data and question coverage that answer engines read. The companies whose models answer the world's questions have homepages those same models struggle to cite.
What the top 10 do that the bottom 10 don't
The gap between the leaders (average 73.0) and the laggards (average 38.8) comes down to a short list of mechanical differences:
| Signal | Top 10 | Bottom 10 |
|---|---|---|
| Structured Data (avg score) | 97/100 | 10/100 |
| Organization schema present | 10 of 10 | 1 of 10 |
| FAQ schema present | 4 of 10 | 0 of 10 |
| Direct-answer block present | 10 of 10 | 5 of 10 |
| Query Coverage (avg score) | 58/100 | 15.5/100 |
Read that table twice: every single top-10 homepage carries Organization schema and opens with a direct answer. In the bottom 10, one homepage has Organization schema and none has FAQ schema. The leaders aren't doing anything exotic — they've done the basics that the laggards skipped.
The differentiator is not budget, brand, or content volume. Datadog and Vercel — companies with world-class engineering — sit in the bottom five because their homepages give an answer engine almost nothing structured to work with.
The one fix that would lift the most homepages
For 37.4% of the companies scored, the single weakest dimension was Structured Data. That makes JSON-LD markup the highest-leverage fix in the entire study: add Organization schema (name, logo, description, sameAs) and a FAQPage block answering the three questions buyers actually ask, and the most common failure mode disappears.
It is also the cheapest fix. Structured data requires no redesign, no new copy deck, and no ranking risk — it is additive markup. A competent developer ships it in an afternoon. Yet 41.8% of the most successful SaaS companies in the world haven't.
The second-cheapest fix addresses the 26.4% of homepages with no direct answer: put one 40–60 word plain-language answer to "what is this product?" near the top of the page, before or directly after the slogan. Our breakdown of how each dimension is scored — including what counts as a direct-answer block — is in the 9-dimension scoring framework.
Methodology
- Engine: AEOCrawler 9-dimension production scorer, rule-based path (EntityClarity v1.27.2). Weights: Answer Extraction 20%, Citation Probability 18%, Structured Data 12%, Entity Authority 12%, Query Coverage 10%, Semantic Coverage 10%, Conciseness 7%, Readability 6%, Content Freshness 5%.
- Run: 2026-07-02, all URLs scored through the same path on the same infrastructure. No mixed model tiers.
- Sample: 100 SaaS companies across 10 categories (10 per category), chosen for recognizability. One URL per company: the primary marketing homepage — the page AI engines most often read for "what is X" queries. 91 of 100 were reachable; the 9 failures are listed above.
- Captured per company: composite score, all nine dimension scores, schema booleans (Organization, FAQPage), and direct-answer-block presence.
- Disclosure: AEOCrawler built the scoring model and authored this study. AEOCrawler's own site was excluded from the ranked sample.
Limitations
- The score predicts citability; it does not measure actual citations. A homepage can score low and still be cited on brand strength — the score measures what the page itself gives an engine to work with. Real-world citation checking is a separate step (AEOCrawler does this with live engine queries; see proactive vs reactive AEO).
- Homepage-only. A company's docs or blog may be far more citable than its homepage. This study deliberately measures the front door.
- Single snapshot. Homepages change weekly; these numbers describe 2026-07-02.
- Extraction noise exists. Scores are computed on the full extracted page text; heavy client-side markup can inflate or deflate text-based dimensions for individual companies. Directional findings are robust across the set; individual ranks a few positions apart should not be over-read.
Frequently Asked Questions
How were the 100 homepages scored?
Every homepage was fetched and scored on the same day through AEOCrawler's production 9-dimension model — the same scorer the product runs on customer pages. Each page gets 0–100 per dimension, and the composite is the weighted sum. No manual adjustments were made to any score.
What is AI visibility?
AI visibility is how likely your content is to be read, understood, and cited by AI answer engines like ChatGPT, Perplexity, Gemini, and Google AI Overviews. It depends on extractable answers, structured data, entity clarity, and question coverage — not on traditional rankings alone. What is AEO? covers the discipline in full.
Why did no homepage reach 80?
Because homepages are built for humans mid-scroll, not engines mid-answer. The two lowest-scoring dimensions — Query Coverage (32.8) and Citation Potential (33.8) — reflect that: most homepages answer one implicit question ("why are we great?") instead of the six an engine gets asked, and offer slogans instead of quotable facts, lists, or tables.
Does a low score mean AI engines never cite the company?
No. Engines also cite third-party sources — reviews, comparisons, documentation — and strong brands get mentioned on reputation alone. A low homepage score means the company is leaving its most-read page unable to speak for it, so engines fall back on what others say instead.
Which SaaS category scored best — and worst?
CRM, sales & support scored best (average 62.8/100), led by Pipedrive at 77.6. Data, analytics & AI scored worst (average 49.7/100) — the companies building AI have the homepages AI can least easily cite.
Can I score my own homepage?
Yes — the same 9-dimension scorer used in this study is free to run on any URL at aeocrawler.com, no signup required. You get the composite score, the per-dimension radar, and the weakest dimensions to fix first.
Score your own homepage — then fix it
Every number in this study came from the scorer that is free on our homepage: paste a URL, get the 9-dimension breakdown. If your score lands where most of this list did, the fix-list is the point — and on Apex, your AI coding agent can read that fix-list over MCP and apply the changes in your editor while you review the diff. Score, fix, re-score. The top of next year's scoreboard is unclaimed.
Update, July 2026: we ran the loop on ourselves
When this study was published, our own landing page scored 61.7 — it would have placed mid-table, right around the 55.5 average. So we did what this study tells you to do: added an on-page FAQ with FAQPage schema, made the key statistics extractable, completed the Organization schema, and wired real freshness signals. Then we re-scored through the same parity path.
61.7 → 84.5. Same scorer, same nine dimensions, no special treatment — the fix-list is the product. You can see the score plotted against this study's distribution on the landing page, and the mechanics of score → fix → re-score in the apply loop.
AEOCrawler remains excluded from the ranked sample above — you don't rank yourself in your own benchmark. This update exists because the study's recommendations should be falsifiable: we applied them, and this is what happened.
