Evidence-first GEO for B2B: AI search visibility you can measure.
For marketing leads at B2B SaaS companies whose buyers now ask AI assistants which vendors to consider. We measure what ChatGPT, Perplexity, Google AI Overviews or AI Mode, and Gemini say about you, then fix only what the evidence supports.
Start with a free AI Visibility Snapshot.
We ask AI assistants the questions your buyers ask and show you what comes back, with a dated picture of today and the first fixes we would make.
Within 3 business days of your reply to our email.
One free diagnostic per company. We take on up to two a week across all services; if this week is full, you are first in next week's queue.
- On the form: your name, work email, website and one sentence on your category.
- By email afterwards: your category in a phrase, 3 competitors and 3 questions your buyers ask before they buy.
- What you send is used only for your Snapshot and deleted after delivery. NDA on request.
- 15 buyer prompts, 5 of them written as your buying-committee roles, run across ChatGPT, Perplexity, Google AI Overviews or AI Mode, and Gemini.
- Whether you are named or recommended, how you are described, who is named instead and which sites the answers cite.
- A one-page scorecard, a 10-minute recording and the top 3 fixes we would make first.
- Answers vary from run to run, so every Snapshot is dated and labelled as a sample.
Not a score. The answers themselves.
In G2's March 2026 survey of 1,076 B2B software buyers and decision-makers, 51% said they now start their research with an AI chatbot more often than with Google (G2, The Answer Economy, published 15 April 2026). So we read what those assistants actually tell your buyers, engine by engine.
The questions
Prompts your buyers type when they choose a vendor, written as the roles on your buying committee rather than as keywords.
Buying committeeVendor choiceThe answers
Each prompt run in ChatGPT, Perplexity, Google AI Overviews or AI Mode, and Gemini. Are you named, recommended and described accurately, and who is named instead?
NamedRecommendedThe sources
Every site the answers cite for your category, and whether it mentions you. That list shows where the work is.
Cited domainsYour own data (Audit and programme)
Search Console's Generative AI report, Bing's AI Performance report, sessions from AI assistants in GA4, and the "How did you hear about us?" field on your forms.
First-partyAnswers change from run to run, so every result we report is dated and treated as a sample of that day, never as a permanent ranking.
Four things we refuse to sell.
You may be offered any of these as GEO. We do not sell them, and the sources below say why. If a proposal you are reading includes one, ask for its measurement plan.
llms.txt as a visibility lever
Google says Google Search ignores llms.txt files, so creating one will "neither harm nor help" your visibility there (Google's AI optimization guide, updated 10 July 2026). A study of nearly 300,000 domains found no correlation between having the file and how often a domain is cited in AI answers (SE Ranking, 7 November 2025). We check crawl and index access for Google, Bing and AI crawlers instead. Our llms.txt guide covers the one case where the file earns its keep: developer documentation.
Guaranteed citations
No one can guarantee a citation. Answers vary from run to run, and Google says no third-party tool has access to its internal ranking or AI systems (Google, updated 10 July 2026). We improve the odds and measure them monthly.
Self-ranked listicles
A "best [category]" list on your own blog with you at number one. In a study of 100 B2B "best software" queries, when a brand's own list was cited in Google's AI Overviews, that brand was left out of the recommendation 69% of the time, in 224 of 323 cases (study published 17 June 2026).
Paid inclusion
Buying spots in lists, reviews or "mentions". Google says seeking inauthentic mentions "isn't as helpful as it might seem" (Google, updated 10 July 2026). We ask lists that publish their criteria to include you on the facts, and we never pay for a place.
Three steps, priced up front.
Each step stands on its own. Start free, and move to the next only if the evidence says it is worth it.
What AI assistants say about you today.
- 15 buyer prompts across all four engines, for you and 3 competitors
- Scorecard, 10-minute recording and top 3 fixes, within 3 business days of your reply
The full baseline and a 90-day plan, before you commit to a programme.
Fixed priceCredited to your first programme month.
- 20 to 25 buyer prompts, including variants written as your buying-committee roles, across all four engines
- Each vendor-choice prompt run three times, because answers vary
- A map of the sites the answers cite in your category
- A scan of your Search Console for long, prompt-shaped queries, so tracker traffic does not inflate your baseline
- Search Console and Bing AI reports set up, with a baseline
- A crawl, index and entity-consistency check
- Content gaps against the sources AI cites, and a 90-day plan
Monthly work in fixed units, reported against your baseline.
Starts at- 1 first-hand asset a month: original data or research only you can publish
- A monthly 30-prompt panel, run as your buying committee
- Search Console AI and Bing citation reporting
- A set number of technical-fix hours agreed in your proposal
- Outreach asking the lists and directories AI answers cite to include you accurately: earned, never paid, inclusion not guaranteed
- Month to month from day one, with 30 days' notice
Prices in USD. Not included in the programme: tool subscriptions, more than one first-hand asset a month, development work beyond the agreed technical-fix hours, and paid placements of any kind. Our published prices, and how we scope the rest, are on our pricing page.
What $1,200 a month buys, in units.
Google's own guide says optimizing for generative AI search "is optimizing for the search experience, and thus still SEO" (Google, updated 10 July 2026). So one programme covers all three: SEO to rank, AEO to win featured snippets and People Also Ask, and GEO to be named in AI answers. More assets or more hours are quoted up front, never assumed.
One first-hand asset a month
Something only you can publish: your own data, a benchmark from your product, a teardown of your method. Google's guide says content people find unique, compelling and useful will likely do more for your presence in generative AI search in the long run than any of its other suggestions (Google, updated 10 July 2026).
A 30-prompt panel, run as your buying committee
A fixed set of 30 prompts, written as the roles who sign off on a purchase like yours, run every month across all four engines. The prompts stay the same, so the trend means something.
Search Console and Bing reporting
Google's Generative AI performance report shows your impressions in AI Overviews and AI Mode by page (launched 3 June 2026, on all sites since 31 August 2026). Bing's AI Performance report counts citations in Copilot and Bing's AI summaries (launched in public preview on 10 February 2026).
Crawl and index access for Google, Bing and AI crawlers
Robots rules, sitemaps, canonical URLs, rendering and internal links, within the technical-fix hours agreed in your proposal. To be a supporting link in AI Overviews or AI Mode, a page must be indexed and eligible for a snippet (Google, updated 10 December 2025), and OpenAI says sites that block its OAI-SearchBot are not shown in ChatGPT search answers (OpenAI, read 23 September 2026).
Entity consistency and schema
The same facts about your company on your site, LinkedIn, directories and review sites, plus structured data for rich results, done within the same technical-fix hours. Google says structured data isn't required for its AI search features (Google, updated 10 July 2026), so we never sell it as a citation lever.
Earned outreach for factual inclusion
Short, factual notes to the lists, directories and publications that your category's AI answers cite, asking for accurate inclusion where they publish their criteria. Earned, never paid, and no inclusion is guaranteed.
Four numbers, every month.
Agreed in month one, measured against that month's baseline and reported the same way every month after.
| Metric | What it counts, and where it comes from |
|---|---|
| Share of panel prompts where you are named | How many of the 30 panel prompts name you, engine by engine, across all four engines. |
| Cited-domain gains | How many of the sites the answers cite for your category now mention you, against the month-one baseline, from the panel's cited-source map. |
| AI-assistant sessions | Visits to your site from AI assistants, in your GA4, with a channel group for AI assistant sources. |
| AI-sourced leads | Leads who pick an AI assistant in the "How did you hear about us?" field on your forms, which we add if it is missing. |
We will not inflate your numbers. Long, prompt-shaped queries show up as Search Console impressions, and many come from AI-visibility trackers and AI search fan-out rather than from buyers. We see them on our own site too, so we strip them out before we quote any impression total.
Our own numbers. In our own checks so far, Omnitics is not named for its own category yet, and every Snapshot we send says so.
Built for one buyer. Honest about the rest.
The marketing lead at a Series A to B B2B SaaS company of roughly 30 to 200 people, selling to a buying committee whose members already ask AI assistants which vendors to consider. You have data or expertise only you can publish, and you want a baseline before a retainer.
Get my free SnapshotNot yet a fit
- You sell mainly to consumers. We work with B2B and B2B2C companies.
- You want guarantees or volume. We promise no ranking or citation, and we ship one first-hand asset a month, not dozens of articles or a page per prompt.
- You have nothing first-hand to publish. The monthly asset is built from your own data, product benchmarks or expertise; without them we would be rewriting what already ranks.
- You need new pipeline this quarter. Search and AI visibility build over months, not weeks. Account-based marketing reaches named buyers sooner.
From the playbook library.
We publish the thinking this service runs on: how to get cited by ChatGPT and Perplexity, how to win the answer box and People Also Ask, whether you need an llms.txt file, the coverage hole AI search opened, and all SEO, AEO and GEO playbooks.
Running account-based marketing as well? When a target account researches the problem, the same work improves the odds that the surfaces they check name you.
Before you ask.
No, and no one can honestly promise it. Answers change from run to run and each engine chooses its own sources. We improve the odds and measure them monthly against your baseline.
With a fixed panel of buyer prompts run across ChatGPT, Perplexity, Google AI Overviews or AI Mode, and Gemini, recording whether you are named, how you are described and which sites are cited.
The AI Visibility Snapshot is free. The AI Visibility Audit is $900, credited to your first programme month. Programmes start at $1,200 a month and run month to month from day one, with 30 days' notice.
Not for AI search visibility. Google's AI optimization guide (updated 10 July 2026) says Google Search ignores it. The file earns its place only on developer documentation; our llms.txt guide has the evidence.
Yes, in what decides whether you can appear. Google's AI Overviews and AI Mode link only to pages indexed and eligible for a snippet (updated 10 December 2025). OpenAI says sites that block OAI-SearchBot are not shown in ChatGPT search answers (read 23 September 2026). We report each engine separately.
Ask ChatGPT about your category. Did you come up?
Get a free AI Visibility Snapshot: how ChatGPT, Perplexity, Google AI Overviews or AI Mode, and Gemini describe you and your competitors today, and the top three fixes.
Free · One per company · Within 3 business days of your reply