AI Visibility Snapshot
From $600, 3 working days
10 buyer prompts across four engines, your homepage and up to eight pages scored on the 22 checks, a one-page readout with the three fixes to make first.
GEO for SaaS
GEO for B2B SaaS, measured against 1,586 software websites
I checked 1,586 software company websites with 11 to 200 staff on 22 October 2026. The median scored 68 out of 100, and 76% of the points lost were in content an engine could quote. The audit shows where your site sits in that spread and which buyer prompts in your category cite someone else, with a list of what to fix first.
AI Search Audit from $2,000, 1 week, fixed price.
Send me your category and two competitors. I run a few of your buyer prompts before the first call, so it starts from your numbers.
By Ronan Liedmeier, fractional head of growth. Last updated .
What 1,586 software websites showed
Software publishers with 11 to 200 staff, founded 2015 or later, in the US and 12 European countries, scored on the same 22 checks as the free tool. Access averaged 93.1% of its points. Answerable content averaged 47.5%.
Median readiness score
68 / 100
Mean 63.9. Only 21.8% scored 80 or more.
Points lost in content
76%
Answerability 46.7% plus Evidence 29.5%. Access caused 5.0%.
Blocked a search bot in robots.txt
0.3%
4 of 1,586 sites. 10.7% failed the text-without-JavaScript check.
Article sites with an FAQ and a comparison table
25.9%
Of 1,303 sites with article pages. Both formats predicted citation in my 209-vendor test.
Commercial sites stating prices
22.7%
309 of 1,361 had a pricing page with amounts.
US and European software sites, side by side
The gap in overall score is not statistically meaningful (z 1.59). The three format gaps below are real (z 4.2 to 6.1).
| US (639 sites) | Europe (947 sites) | |
|---|---|---|
| Mean score | 64.9 | 63.3 |
| FAQ on article pages | 62.3% | 45.0% |
| Comparison table on article pages | 42.7% | 27.2% |
| Pricing page with amounts | 28.5% | 18.8% |
Method, exclusions and limits are in how ready software company websites are for AI search. A readiness score and a citation are different measures, and the audit runs both.
The prompts SaaS buyers type
A SaaS shortlist gets built from comparison and price questions. The panel uses shapes like these, filled in with your category, your buyers' company size and the competitors your sales team hears about:
- best {category} software for a {size}-person {industry} company
- {your product} vs {competitor}
- {competitor} alternatives
- how much does {category} software cost per seat
- {category} tool that integrates with {system}
- is {your product} GDPR compliant
Each of those questions gets answered from somebody's page. If your comparison page has no table and your pricing page has no numbers, the engine quotes a review site or a competitor that does.
Who this is for
- A B2B software company in the US or Europe, roughly 11 to 200 staff, whose buyers shortlist tools in ChatGPT, Perplexity or Google before they talk to sales. I run the prompts from the country your buyers are in.
- Someone on your side can change the site in the next quarter: a marketer, a developer or an agency.
Who it is not for
- You want a guaranteed number of citations. The engines do not publish their selection rules and nobody can promise a slot.
- Nobody can ship changes for a quarter. The fix list would sit unread.
- You want the pages written for you. That is the Pipeline Build or a Fractional Head of Growth engagement.
What the audit week covers
- 1
Prompt panel in your category
Before the audit starts I write the prompts your buyers type, with company size, country and the competitor they are comparing you to, and run each one in a fresh thread on ChatGPT, Perplexity, Gemini and Google AI Overviews. For every answer I log whether your domain was cited, whether your brand was named, and which domain took the slot.
- 2
Crawl of your site
The 22 checks from the free tool, on your homepage and at least eight content pages, plus bot probes as OAI-SearchBot, PerplexityBot, GPTBot and ClaudeBot. This is where JavaScript-only pages, CDN bot rules and snippet limits show up.
- 3
Competing domains, question by question
The list of who owns each prompt and in what format: a comparison page, a pricing page, a review site, a Reddit thread. For SaaS categories this list is usually the content plan.
- 4
Fix list and walkthrough
Blockers first, then the article-level checks that predicted citation in my 209-vendor test, then everything else. A 30-minute call to go through it with whoever runs the site.
I order it from my test of 209 vendors named in Perplexity answers: the checks that separated cited from uncited vendors in the same answer were an FAQ section, a named author, article depth, comparison tables and visible dates. Organization schema showed no effect, so it sits at the bottom of the fix list.
Which engagement fits
The same three engagements as on the AI visibility page, with the prompt panel written for a software category.
AI Search Audit
From $2,000, 1 week
20 buyer prompts, a full readiness audit of your homepage and at least eight content pages, a fix list ordered blockers first, a content plan and a 30-minute walkthrough.
AI Visibility Monitoring
From $400 a month
Up to 60 prompts frozen for a quarter, run weekly across four engines, with a monthly report on what moved and which two or three pages to write next.
Prices are in USD; European clients are invoiced in EUR at the same figures. Fixed projects are invoiced 50% on signature and 50% at handover. No hourly billing. Full inclusions are on the AI visibility page.
Common questions
What is GEO for a SaaS company?
GEO, generative engine optimization, is the work of making a site readable and quotable for AI answer engines such as ChatGPT search, Perplexity and Google AI Overviews. For a SaaS company it mostly comes down to the questions buyers ask while shortlisting: which tool for a given company size, how one product compares to another, what the alternatives are and what it costs.
Where do software company websites usually fall short?
In content an engine can quote. In my check of 1,586 software company websites on 22 October 2026, 76% of the points lost were in answerable content and evidence, while access caused 5.0%. Of the 1,303 sites with article pages, 51.9% had an FAQ and 33.4% had a comparison table.
Our site is built with React, Lovable or Webflow. Is that a problem?
Only if the page text is missing from the HTML the server sends. In the 1,586-site check, 170 sites (10.7%) failed the text-without-JavaScript check. Vercel's December 2024 crawler analysis found none of the major AI crawlers rendered JavaScript, Google's Gemini aside, so a client-rendered page can rank in Google and still be blank to ChatGPT. The audit runs this check on every sampled page.
Should we publish our prices?
If you have list prices, yes. A buyer asking an engine what a tool costs gets an answer from somebody's page, and only 22.7% of the 1,361 commercial software sites in my check stated amounts on a pricing page. In the US the share was 28.5%, in Europe 18.8%.
Do you promise we will be cited?
No. I promise a measured baseline in your category, which is the number you can hold me to, and a fix list ordered by what blocks citation before what improves it. My own site's baseline was zero citations on Perplexity, Gemini and Google AI Overviews on 16 October 2026, and the day-90 target is published in the Insights section.
What does it cost?
The AI Visibility Snapshot is from $600 and takes 3 working days. The AI Search Audit is from $2,000 and takes 1 week. AI Visibility Monitoring is from $400 a month. European clients are invoiced in EUR at the same figures.
Send me your category and two competitors. I run a few of your buyer prompts before the first call, so it starts from your numbers.