We get B2B software onto the pages AI quotes.
The hard part isn’t knowing you’re missing. It’s getting onto those pages. Every month we send 6–10 researched pitches for your product to the roundups, comparisons and review platforms the engines cite in your category, plus 2 publish-ready pages for your own site.
Free breakdown first. If the engines already name you every run, we’ll say so and it’ll have cost you one email.
Get your free breakdown
That didn’t go through, please check the required fields and send again.
Blueprint $2,900 · pricing published · a client, measured
measured 2026-07-16 · 71 domains recorded · contract management software
Top 6 of 71. 1 more tie at 8/32. the full source list →Three things that decide whether AI names your product
of the sources ChatGPT cited when recommending B2B software were the vendors’ own pages. On Google AI Overviews, 12%.
of named vendors showed up in half their identical reruns or fewer. Ask once and you get one draw.
of cited sources were shared between engines. They name similar vendors and read different pages to get there, so the target list is built per engine.
CrediGeo study · 1,040 answers · 4 engines · 8 B2B hiring-software categories · 2026-07-21
We keep measuring. The AI Visibility Index publishes 12 B2B software categories, re-measured monthly, with every one of its 574 answers archived and readable.
The answer your buyers see is assembled from pages you don’t own
51% of B2B software buyers now start their research with an AI chatbot more often than with Google, up from 29% a year earlier. The model answers with a shortlist. Ask the same buyer question enough times and the same pages keep supplying that shortlist. Here is one real category, measured on July 16, 2026: who the engines named, run by run.
G2, 2026 AI Search Insight Report, 1,076 B2B decision-makers, fielded March 2026. G2 sells into this market.
Each square is one run, left to right.
Icertis flipped between identical runs, same question, same day.
measured 2026-07-16 · n=5 runs · ChatGPT · contract management software
how we run this →AI hands your buyers a shortlist
A few named vendors. And many buyers never scroll past them to a Google result.
You can’t see the answers you’re losing
Private chats, no impression count, nothing in GA or your CRM.
It’s built from a handful of pages
Mostly pages someone else controls. Getting onto them is the work; your Google rank doesn’t decide it.
The full measurement, vendor by vendor → · the 1,040-answer study behind the numbers above →
What a month of the work contains
Two layers, because the source list has two layers: the pages you own, and the pages you don’t. You see everything before it moves, and everything carries a date.
-
The source map, before anything is written
Which pages your category’s answers are actually built from, which of them name competitors while they don’t name you, and the order worth working. That is the Blueprint.
-
6–10 researched pitches a month
We find who owns the roundups, comparisons and review platforms the engines cite, and write each of them a short evidenced note proposing your inclusion. We send them and we follow up. You read every note before it goes.
-
2 publish-ready pages a month
Answer-shaped pages for your own site: comparisons, alternatives, plain-language product facts. Written so an assistant can lift a paragraph and have it still be correct.
-
No paid placements, no fake reviews
No paid placements. No fake reviews, no invented accounts, no comment-section astroturf. If a page owner won’t have you, that goes in the log and we move to the next name.
Every published number ships with its archived raw answers and the exact prompts. The full method →
the same face you’ll meet on the walkthrough call
Frank, founder
One person does the work, and you deal with him the whole way
CrediGeo is just me. No team, no account managers, no handoffs.
I’ve spent the last seven years in data engineering, building pipelines and web-data products where a wrong number gets caught quickly. I measure AI visibility the same way. Ask the same question many times, log every answer, and claim nothing the log doesn’t show.
If we work together, you deal with me the whole way. I write the pitches, I send them, and I’m on every call. The Vokaro case below is my work.
The one client we’ve done this for
We wrote Vokaro’s pages. Google’s AI quotes them back.
Ask Google’s AI Mode the questions Vokaro’s buyers actually type, in its own market and language, and it names Vokaro in 8 of 9 answers. In 8 of those, vokaro.net is one of the sources the answer was built from, the same runs, not a different set. The two URLs it cites are the use-case page and the category comparison we wrote. Nobody paid for a placement.
| Domain | Cited in |
|---|---|
| placetel.de | 8/9 |
| youtube.com | 8/9 |
| vokaro.net their own site | 8/9 |
| fonio.ai | 6/9 |
| ionos.de | 6/9 |
We hold no dated placement on any of those third-party pages, and that is exactly what a sprint’s 6–10 researched pitches a month go after.
The other half, same client, same day. ChatGPT named it just 2 of 15 times in that market. And in the US market, which this engagement never targeted, it was 0 of 12 on AI Mode and 0 of 20 on ChatGPT. One engine reads the open web far more readily than the other, and a number from one market and language does not carry to another. Yours would be measured in yours.
These are Vokaro’s own analytics. Over the whole window, 5 different assistants sent visits, up from one at launch. Read these as a floor, not a total, because assistant apps often send no referrer at all. The bars sum to the 88 sessions published in full on the case study.
measured 2026-07-17 · DACH · de · analytics 2025-12-19–2026-07-17
What the Vokaro data establishes, and where it stops
Most companies we measure should not hire us.
One client, one market, one language. vokaro.net is that client’s own site, the pages we wrote. That is the half we can show you a number for. We have no dated third-party placement in that dataset, and we won’t imply one until we do.
- Who it’s for
- Competitors named while you aren’t, or a position that flips run to run.
- Who it isn’t for
- If the engines already name you consistently across runs, the report says so and we don’t pitch the sprint.
- Why we run it many times
- In the run card above, Icertis was named in 2 of 5 identical ChatGPT runs. One check would have called it present or absent depending which run you caught. That argues for repeating the question, not that this category is unusually unstable. Measured against pure chance at these rates, it isn’t.
We don’t promise placements or rankings. No page owner is obliged to say yes, and a spot in an AI answer can’t be bought.
A tracker tells you the number. It doesn’t send the pitches.
| Category | A visibility tracker | CrediGeo |
|---|---|---|
| What you get | A score, and prompts you are missing from | The source list, then the pitches and pages that go after it |
| Who does the outreach | You do, or nobody does | 6–10 researched pitches a month, follow-ups included |
| Proof of work | A number in a dashboard moved | Every pitch, page and raw answer dated and archived |
| Terms | A seat, monthly, indefinitely | A fixed 90-day sprint, billed monthly, cancel any month |
Start with your own source list. Your buyers’ questions, both engines, and the pages those answers were built from.
- Engines
- ChatGPT and Google AI Mode
- Runs
- 3 ChatGPT and 2 AI Mode per question, each a fresh session
- You get
- Every answer, who was named in it, and the exact prompts
- Arrives
- Within 2 business days, from a person
- Cost
- None, and no call required
Get your free breakdown
A run is one question, asked once, in a clean session. We ask your buyers’ questions on ChatGPT and on Google’s AI Mode, more than once on each, then send you who got named in every answer with the exact prompts.
That didn’t go through, please check the required fields and send again.
Published pricing. The Blueprint picks the targets; the sprint works them.
- The Blueprint: the source map, which pages your category’s answers are built from, and where competitors are named while you aren’t, plus the 90-day plan built from it
- The sprint: 90 days of that plan executed. 6–10 researched pitches and 2 publish-ready pages a month, with the same measurement run on day 0 and day 90
- Published in writing. No annual lock-in, no “call for pricing”
Whether that is worth it is arithmetic you can do, and we will not do it for you. Take your deal size, your win rate, and how many evaluations start with a shortlist you are not on. We cannot measure the deals that never started, and nobody can, so we do not publish a figure for it. What we can tell you is what it costs and exactly what arrives.
What buyers ask before the first call
- Do you do the outreach, or just tell me where I’m missing?
- We do it. 6–10 researched pitches a month to the pages the engines cite in your category, every one shown to you before it goes, plus 2 publish-ready pages for your own site. No paid placements and no fake accounts, and no page owner is obliged to say yes. What a sprint month contains →
- What happens when a site says no, or says nothing?
- Most won’t reply. That is why the volume is 6–10 a month and why every pitch and every reply is logged with a date. You read the real hit rate instead of being told one.
- Can you show this working on a real company?
- Yes, on one. Vokaro is the only client we’ve done this for, and the whole engagement is published: the prompts, the numbers market by market, and every raw answer. Read the case study →
- Why not just use a free AI-visibility checker?
- Do. And our free breakdown answers the same question run by run, for nothing. What no checker tells you is which pages those answers were assembled from and who controls each one. None of them does the work of getting you onto those pages either.
- Do B2B buyers actually use AI to choose software?
- 51% of B2B software buyers now start research with an AI chatbot more often than Google, up from 29% a year earlier (G2, 1,076 B2B decision-makers, fielded March 2026). The model answers with a short list of vendors. Most companies have never checked whether they’re on it. The source, and who funded it →
- Who actually does the work?
- Frank, the founder. CrediGeo is a one-person company: he runs the measurement, does the research and the outreach, and takes every call himself. No account team.
- Is this just SEO or AEO with a new label?
- No. SEO ranks a page in a list of links; this measures whether the AI names your product at all, a different surface, read from different sources. The full comparison →
Get your free breakdown
A run is one question, asked once, in a clean session. We ask your buyers’ questions on ChatGPT and on Google’s AI Mode, more than once on each, then send you who got named in every answer with the exact prompts.
That didn’t go through, please check the required fields and send again.
What arrives, exactly
- Engines
- ChatGPT and Google AI Mode
- Runs
- 3 ChatGPT and 2 AI Mode per question, each a fresh session
- You get
- Every answer, who was named in it, and the exact prompts
- Arrives
- Within 2 business days, from a person
- Cost
- None, and no call required
Prefer to talk first? Book a walkthrough (opens in a new tab). Your category on screen, no deck. Or read the real sample report.