Key takeaways
- We ran our live AI visibility audit tool against 35 B2B SaaS companies and could scan 30. Average visibility across those 30 was 67.3% over 150 live, grounded buyer queries. The median company scored 80%.
- The spread is wide. Eight companies scored a perfect 100%, eight scored 40% or less, and two (folk and June) were not named in a single query.
- Problem-first queries were the weak spot: companies were named 50% of the time, against 83% for comparison queries. All eight perfect scorers were named on the problem-first query. Only 7 of the other 22 companies were.
- Citations came from a long, fragmented tail: 1,184 citations across 663 distinct domains. The most-cited domain, G2, accounted for 21 of them (1.8%). A company's own website made up about 3% of everything cited.
- Five of the 35 companies could not be scanned because their sites showed our crawler a bot-verification page. That is a finding in its own right.
- This audit is of established companies. It says nothing about how brand-new SaaS sites perform, and it likely flatters the average for a company with no market presence yet.
Why we ran a SaaS-only audit
In our pilot audit of 29 B2B companies across six industries, B2B SaaS looked like the best-performing sector. We said at the time that the sample was only two companies, which is far too few to conclude anything. This is the proper version: the same tool, the same method, thirty SaaS companies.
The question is simple. When a software buyer asks an AI assistant for options, without naming any company, who gets named? And does being a real, funded, well-known SaaS company guarantee you are one of them?
67.3%
average AI visibility across 30 real B2B SaaS companies, measured by live grounded search on 150 buyer queries
Deligatr SaaS AI Visibility Audit, October 3, 2026
How the audit works
The tool reads each company's homepage, classifies what it sells in the words a buyer would use, writes five buyer-intent queries (one local-style, two category, one problem-first, one comparison) and runs each as a live, grounded AI search with the company's own name left out. It is one run, five queries per company, on one day. AI answers change between runs, so read every score as a snapshot. The full method is in our audit documentation, and all twelve industry audits are on our AI visibility research page.
A company is either in the answer or it is not.
The distribution: a long tail at the bottom
An average of 67.3% hides how uneven the results are.
Companies by AI visibility score
30 B2B SaaS companies, each tested on 5 live grounded buyer queries. Number of companies at each score.
Deligatr SaaS AI Visibility Audit, October 3, 2026
Eighteen of the thirty companies were named in at least four of five queries. Eight were named in two or fewer. The two zeros, folk (CRM) and June (product analytics), both sit in categories where a few older names dominate the answers: Pipedrive and HubSpot for CRM, Amplitude, Mixpanel and Pendo for product analytics.
Being a real, funded product did not protect anyone. PostHog, which is widely known among developers, scored 20%. We come back to why below, because it is instructive.
The query that separates winners from the rest
Not every query is equally hard. Here is how often a company was named, by query type.
How often a company was named, by query type
Share of queries in which the company was named. 30 SaaS companies, 150 queries.
Deligatr SaaS AI Visibility Audit, October 3, 2026
The problem-first query is the one that sorts the field. All eight companies that scored 100% were named on it. Of the other 22 companies, only seven were.
SaaS companies do better here than the businesses in our cross-industry audit, where problem-first queries named a business only 21% of the time. They still lose 33 points between the comparison query and the problem-first query. That is the top of the funnel: the buyer who has a pain and no vendor in mind yet. Most SaaS sites have a feature page and a pricing page, and very little that answers the problem in the buyer's own words.
A note on the "local-style" bucket. The tool generates one location-flavored query per company. For software that is often a "near me" or "for my region" phrasing that a real SaaS buyer would not use, so treat that row as the least meaningful of the four.
Who gets named instead, and where the answers come from
When a company was not named, AI assistants named someone else. The same names recurred across categories: Pendo and Auth0 each appeared as a named alternative in 10 of the 150 answers, with HubSpot, Apollo.io, Appcues, Intercom and Zendesk in nine each. Several cold email tools (Instantly, lemlist, Smartlead) named each other as the top alternative, which shows how tightly a category's leaders cluster in the answers.
The sources behind those answers were scattered. Across the 150 queries we recorded 1,184 citations spread over 663 distinct domains. The most-cited domain was G2 with 21 citations, 1.8% of the total. The most-cited domains included review sites, roundup blogs and vendor comparison pages.
3.0%
of all citations pointed to the company's own domain (36 of 1,184), counting citations from the company's own domain and its subdomains
Deligatr SaaS AI Visibility Audit, October 3, 2026
That number matters more than it looks. Your own site is a small share of what AI assistants read about you. The rest is third-party pages: reviews, roundups, comparison posts and community threads. A SaaS company with a good site and no mentions elsewhere has covered only a fraction of the ground. The upside is that the citation pool is open. No single domain dominates it.
The five we could not scan
Five of the 35 companies we tried (Clay, Drata, Heap, Personio and WorkOS) could not be audited. Our crawler was shown a bot-verification page, either a Cloudflare "Just a moment" challenge or a Vercel security checkpoint, or an error, instead of the homepage.
That is 14% of the sample. Bot protection is sensible for a site that handles logins and payments. What we cannot tell from outside is whether the same rules also stop the crawlers AI assistants use to read your pages. If they do, the content is not being read, however good it is. It is worth checking what your own security layer does to automated visitors.
What a low score tells you
Look at PostHog. Its homepage describes a broad set of products: analytics, session replay, feature flags, error tracking. The classifier read that and generated queries about bug tracking and code debugging, categories where Sentry and others are the established answer. PostHog was named on one of five queries.
That is a measurement artifact, and also a real lesson. When a homepage describes many things, AI assistants have to guess which category you belong in, and you are measured against whoever dominates each guess. Several of the companies that scored 100% sit in a single, clearly named category (subscription analytics, customer feedback, product management, compliance automation). We did not test whether homepage clarity causes the difference, but it is the first thing we would check.
What to do with this, if you run a SaaS company
- Check whether you can be read. Search
site:yourdomain.com, and test what your security layer shows a plain automated visitor. A page nothing can read cannot be cited. - Say what you are, in one sentence, near the top of the homepage. Name the category in the words a buyer uses. It is one of the cheapest fixes on this list.
- Write for the problem-first query. Build pages that answer the pain before the buyer knows your category: how to diagnose it, what the options are, what to check. This is where our data shows the largest gap.
- Get mentioned somewhere that is not your site. With only about 3% of citations pointing at the company's own domain, review sites, roundups and community threads do much of the work. Earn those mentions openly.
- Re-test on a schedule. AI answers shift, so a one-time score tells you little.
If you want this done as a program rather than a to-do list, see our SaaS SEO and AEO service. It covers the technical, conversion and full-funnel content work, written by people and led by data like this. If your buyers also need to be found proactively, Full Hybrid adds signal-based outbound in the same CRM.
Book a free consultation and we will run your own constructed questions against live assistants and show you who is named instead of your company. Inbound is $2,500 a month on monthly billing.
Book a consultationEvery company we audited
Company | Visibility | Appeared on | Top rival named instead |
|---|---|---|---|
| Baremetrics | 100% | Local, Category x2, Problem, Comparison | ProfitWell |
| Canny | 100% | Local, Category x2, Problem, Comparison | Productboard |
| Chargebee | 100% | Local, Category x2, Problem, Comparison | Stripe Billing |
| Gong | 100% | Local, Category x2, Problem, Comparison | Clari |
| Paddle | 100% | Local, Category x2, Problem, Comparison | Chargebee |
| Productboard | 100% | Local, Category x2, Problem, Comparison | Aha! |
| Remote | 100% | Local, Category x2, Problem, Comparison | Deel |
| Secureframe | 100% | Local, Category x2, Problem, Comparison | Vanta |
| Chameleon | 80% | Local, Category x2, Comparison | Appcues |
| Clerk | 80% | Local, Category, Problem, Comparison | Auth0 |
| Help Scout | 80% | Local, Category, Problem, Comparison | Intercom |
| Instantly | 80% | Category x2, Problem, Comparison | Lemlist |
| lemlist | 80% | Local, Category x2, Comparison | Smartlead |
| Loops | 80% | Local, Category, Problem, Comparison | ActiveCampaign |
| Retool | 80% | Local, Category, Problem, Comparison | Appsmith |
| Smartlead | 80% | Local, Category x2, Comparison | Instantly |
| Tally | 80% | Local, Category x2, Comparison | Typeform |
| Userflow | 80% | Local, Category, Problem, Comparison | Userpilot |
| Cal.com | 60% | Category x2, Comparison | Calendly |
| Customer.io | 60% | Local, Category, Comparison | Braze |
| Front | 60% | Category, Problem, Comparison | Zendesk |
| Stytch | 60% | Category x2, Comparison | Auth0 |
| Attio | 40% | Category, Comparison | Pipedrive |
| Grain | 40% | Category, Comparison | Fathom |
| Plain | 40% | Category, Comparison | Zendesk |
| PostHog | 20% | Local | Sentry |
| Sendbird | 20% | Category | Ada |
| Stigg | 20% | Category | Schematic |
| folk | 0% | None | Pipedrive |
| June | 0% | None | Amplitude |
Not scanned (bot-verification page or error): Clay, Drata, Heap, Personio, WorkOS.
Check your own number
Every company above was run through the same tool below. It takes about ninety seconds, and your own name is never in the query.
Do AI assistants recommend you?
Five buyer-style queries for your category, run through live AI search, show who gets named instead of you. Takes about ninety seconds.
Free. The full report arrives by email so the tab can close.
Methodology and honest limitations
What we did. On October 3, 2026 we ran the same live audit tool that powers our free checker against 35 B2B SaaS companies: one run, five queries per company, visibility measurement only, not the conversion audit. The step-by-step method is in our audit documentation. A company that disputes its result can ask us to re-run it.
The sample. We chose real, identifiable SaaS companies across customer support, product analytics, billing, security, sales and marketing tools, with a mix of well-known and smaller names. It is a convenience sample of established companies, not a random one and not a market census. It likely flatters the average, because companies with no market presence yet are not in it. One of the five queries is a "local-style" query, which has little meaning for software, so it dilutes comparisons with other industries.
What it does not cover. Only the homepage was read, so a company with a thin homepage and strong inner pages may be undercounted. Five queries per company is a small sample of real buyer language. The classifier can misread a homepage that describes many products, as with PostHog. Results come from one run on one day, and AI answers change over time, so a rerun would not return identical numbers.
On the earlier audit. The 29-company audit included two SaaS companies. This audit replaces that figure for anything about SaaS specifically.
FAQ
What does AI visibility mean in this audit?
Whether a SaaS company gets named as an option when an AI assistant is asked a real buyer question in its category, without the company's own name in the query. We measured it with live, grounded AI search. A company either appeared in the answer or it did not.
How visible are B2B SaaS companies in AI answers?
Of the 30 companies we could scan, the average visibility was 67.3% across 150 buyer queries. The median company was named in four of five queries. Eight companies were named in two queries or fewer, and two were not named at all.
Which kind of query do SaaS companies lose most often?
Problem-first queries, where the buyer describes a pain and has not named a category yet. Companies were named 50% of the time on those, against 83% on comparison queries. All eight companies that scored 100% were named on the problem-first query, while only 7 of the other 22 were.
Does the audit say anything about early-stage SaaS companies?
No. We audited established companies with real market presence, so this sample likely flatters the average. We did not test brand-new or very small sites.
Why could five companies not be scanned?
Clay, Drata, Heap, Personio and WorkOS returned a bot-verification page (a Cloudflare "Just a moment" challenge or a Vercel security checkpoint) or an error to our audit crawler. Sites that handle logins and payments often run strict protection, and we cannot tell from the outside whether it also stops the crawlers AI assistants use.
Is this the same audit as the 29-company one?
It uses the same tool, the same prompts and the same five-query method, run on a different set: 30 B2B SaaS companies instead of six mixed industries. The earlier audit had only two SaaS companies, which is too few to say anything about SaaS on its own.
