Asked over and over
One lucky prompt proves nothing. Every buyer question goes to every engine several times, about a hundred answers in all, and I count who gets named.
I ask ChatGPT, Claude, Gemini and Perplexity what your buyers ask, about a hundred times, and show you who they name instead of you.
Most businesses find out they're invisible to AI by accident. This is how you find out on purpose.
RevPack: named in 20% of AI answers
Real businesses, real answers from ChatGPT, Claude, Gemini and Perplexity. Every card opens the full report.
Straight from the audits: real answers, named and not
“Who are the best RevOps consultancies for B2B SaaS companies doing $5M ARR?”
For a B2B SaaS company around $5M ARR, the best RevOps consultancies are usually the ones that explicitly work with growth-stage SaaS, can handle hands-on CRM + GTM systems work, and aren’t positioned only for enterprise accounts. Based on current public profiles and…
“What career coach specializes in helping SaaS sales reps land new jobs?”
One strong fit is Amanda Rice — she describes herself as a former SaaS sales leader turned career coach and says she helps people navigate SaaS sales career transitions and land offers in the competitive job market.…
“Who are the top check valve manufacturers that work with new space launch vehicle companies?”
Here are some of the most established check-valve manufacturers that appear to work with new space / commercial launch vehicle companies: • The Lee Company — says launch vehicles typically incorporate hundreds of its check valves, and it specifically serves the space…
“Is there a fractional CMO near Blackburn, England who works with founder-led B2B companies?”
Yes! Based on the search results, Salim Sayed is a Fractional CMO based in Blackburn, England who specifically works with founder-led B2B companies. According to his profile, Salim: • Specializes in founder-led B2B businesses…
“What branding agencies specialize in working with nonprofits and mission-driven organizations?”
Here are some branding agencies that explicitly specialize in nonprofits and mission-driven organizations: • Constructive — A branding and interactive design firm focused on nonprofits, philanthropies, and mission-driven businesses; strong emphasis on brand strategy…
“What are the best water dynamics training products and how do they rank and compare?”
I’m not sure what you mean by “water dynamics training products” because that phrase can refer to a few very different categories: 1. Fitness/training tools that use moving water for instability/resistance, like the Tidal Tank. Its site explicitly markets “water…
“What are the best industrial mixer manufacturers for custom agitators?”
If you need custom agitators, the strongest manufacturers are usually the ones that do application engineering, not just catalog mixers. Based on current manufacturer positioning, good places to start are: • Dynamix Agitators — explicitly says it custom engineers…
“AI workflow audit consultant Miami”
Several Miami-based firms offer workflow-focused AI audits. Eloré Consulting runs what they call a Workflow Audit — a short, usually single-session engagement to watch how a team actually works and name the highest-leverage automation before committing to a larger…
“What backend should I use for a SaaS web app”
For most SaaS web apps, a managed Postgres platform like Supabase or Neon plus a framework like Next.js covers 90% of needs. Supabase adds auth, storage, and realtime on top of the database, which shortens time-to-first-customer significantly. If your team prefers…
Plenty of tools will hand you a score. This one shows its working.
One lucky prompt proves nothing. Every buyer question goes to every engine several times, about a hundred answers in all, and I count who gets named.
Every rate comes with its 95% range. 1 in 5 from 25 answers is a weaker claim than 1 in 5 from 250.
Being mentioned in an answer and being the source it links to are two different wins. They're counted separately.
Sampled through the engines' APIs the way a buyer would use them. Every report names the exact models.
You see the real answers word for word. Quotes from Reddit, X and the web are linked, and marked when the link was checked live.
I review every audit before it goes out. If a question doesn't sound like something your buyer would type, I rewrite it and the report says so.
You can rank first on Google and still lose every AI answer to a directory that isn't you.
Ivan Karaman
Writer-engineer
I build space simulators for a living. I love physics, and search is the closest thing the web has to it: a system you can't see, that still gives up its rules if you measure it patiently.
AI answers made it more interesting. Nobody can tell you why ChatGPT names one business and not the next, so I built a pipeline that asks a hundred times over, and counts.
Questions, a custom job, or a sim for your company:
Video coming soon
$0 by request
I run a few a week, for sites I think I can help. Email me your site and what AI should recommend you for.
$297 once
I write the fixes for your site, not a to-do list.
Not worth it to you? Full refund. Agencies charge $1,500+ for this.
Quoted per project
When the fix needs hands: the pages themselves, a site rebuild, web and app builds, or 3D physics sims for deep tech.
“Excellent job developing our AEO/GEO and AI search visibility tool”
Scott Melnyk Founder, 360 Media Consulting“Built our SEO pages end to end, so we spin up new ones without rebuilding”
Dustin Jones Founder, uBuyFirst“Hit $960k/mo from $200k average”
Tommy Joiner Co-Founder, WordAgents“One of the smartest product-minded devs I've worked with”
Enoc Tagro Founder, OutOfReach3D“Probably the smartest copywriter. Also a very neat SaaS founder.”
Dr Jameson Day Investor, SummaUp“Working with Ivan is a JOY. Always love what he brings to the table”
Kees van Velzen Co-Founder, Learnforce“A level above & beyond anything I could have hoped for”
David Thomson Founder, SuadaFair question, there are plenty. This one runs on ivankaraman.com under my real name, shows the raw answers and the method behind every number, and I don't send reports nobody asked for. If the fix isn't worth it to you, I refund it.
More buyers now ask ChatGPT or Perplexity for a shortlist before they ever search. Those answers are built from a handful of sources, and ranking on Google doesn’t guarantee you’re one of them. The audit shows you whether you are.
Most tools run a prompt once and give you a score. This asks each question many times across four engines, shows the range of uncertainty, keeps mentions and citations apart, and shows you the actual answers. Then a person reads it.
Sometimes. I run a few free audits a week for sites I think I can help. Email me your site and what you want to be recommended for. If I pass, I'll say why.
A deeper re-run of your audit, with more buyer questions; llms.txt and structured data written for your site; Ranked fixes with exact steps for your platform; A brief for the page AI is most likely to cite; A Loom walking you through all of it; A re-test once your changes are live. Delivered within 48 hours, no calls needed. The re-test runs within 90 days of delivery.
Then the audit says so, and you’ll see which engines and which questions. That’s worth knowing too, and you don’t need to buy anything.
Only if you say yes. You get a private, unindexed link first. Published audits live on this site, which is how other people find out this exists.