This image came to me mid-conversation, during an interview with Thomas Pontiroli, a journalist at Les Échos, about our Observatory. I was trying to explain why messaging decides B2B sales, and the parallel imposed itself. The best ideas rarely come from sitting alone in front of a screen; this one was born from a human exchange, and I am developing it here with the help of my favourite AI, who will recognise itself. The article is online at Les Échos (in French, subscribers only).
A CIO has a problem to solve. On the same afternoon, he does two things. He calls his Gartner analyst, and he opens Perplexity. Twice the same question, almost word for word: « what is the best solution for my problem? »
On one side, a human who costs tens of thousands of euros a year and answers by appointment. On the other, a free machine, available at 3 a.m. And yet the two do the same job, turning an unreadable market into a usable answer.
What is not in that answer does not exist.
What actually happens during an analyst inquiry?
People often picture an inquiry as a directory lookup. The client asks, the analyst produces three names, thank you, goodbye. It does not work like that, and that is precisely why the message matters so much.
An inquiry lasts thirty minutes. The client lays out their context, their industry, their size, their constraints, what they have already tried. On the other end, the analyst draws on what they have accumulated over hundreds of comparable conversations. Names, yes, but above all what sets these solutions apart from each other, and what similar clients concretely gained by deploying them. « This one integrates better with your environment. That one, a manufacturer your size rolled it out last year, they cut their time-to-market by 30%. The third is solid but its European support is still thin. »
That is the substance of an inquiry. Argued comparisons, benefits observed among real clients, and the caveats that come with them.
As for where the analyst gets all this, the answer is less mysterious than you might think: from vendor briefings, from their websites, from their case studies, from what the market says. If your differentiator is written nowhere, if your customer benefits sit locked in a sales deck, the analyst has nothing to redistribute. They cannot invent your story for you. They might mention you in passing, without relief. Most of the time, they will not mention you at all.
Gartner claims more than 500,000 client interactions per year. A single analyst covering a hot category handles several hundred on their own. Each of those conversations is a chance to be recommended with your differentiators and your proof, or to miss your turn without ever knowing it.
You are not bad. You are absent.
Let analyst relations managers read this carefully: briefings remain essential. That is where the closeness between a vendor and its analysts is built, and that closeness is cultivated over time, as anyone in the trade will tell you. But if the starting message is not right, every briefing is spent re-explaining who you are instead of moving the relationship forward. The closeness will come anyway. It will just take much longer.
The AI does the same job, at a different scale
The same buyer’s second reflex is more recent. According to The Answer Economy, a study published by G2 in April 2026 among 1,076 B2B software buyers, 51% now start their research in an AI chatbot rather than a search engine. A year earlier, they were 29%. In twelve months the market’s front door moved, and I am not sure many marketing teams have measured what that implies.
Look at what the machine answers when asked the CIO’s question. Not a list of links. A comparative synthesis, with the options, their respective strengths, the benefits users report, the caveats. Structurally, it is an inquiry. No appointment, no subscription, served dozens of times a day in your category alone.
The same study, incidentally, ranks chatbots as the number one influence on short-lists, at 54%, ahead of review sites and ahead of vendor websites themselves. That last point deserves a moment. Your own website weighs less in the short-list than what an AI says about you.
And the entry test does not change from one reader to the other. An AI does not know the intrinsic quality of your product, it has no way to measure it. It recommends what it can extract and compare. Give it a differentiator written down in black and white and quantified benefits, and it has something to cite. Give it a generic message, and it moves on to the next one, as the analyst did before it.
The exclusion is silent, and that is the worst part
Nobody warns you that you failed the test.
The analyst does not call back the vendors they did not mention. The AI even less. The deal never shows up in your CRM because, for you, it never existed.
The G2 figures give an idea of what plays out in that silence: 69% of buyers ended up choosing a different vendor than the one they had planned, based on chatbot answers. One in three bought from a vendor they had never heard of before the conversation. In other words, established brands lose deals they will never know they played, to unknowns with a better story.
How many companies pass the test today? We measured it, that is precisely what our Observatory is for. 369 French post-funding startups, audited on their public data only, meaning what an analyst or an AI sees before any contact. Average clarity score: 5.33 out of 10. Three startups out of four below the critical threshold. Readiness for AI reading: 18% of its potential.
Raising money changes nothing, by the way. The correlation between amount raised and message clarity is close to zero across our corpus. A Series B with a generic message remains a Series B with a generic message. We would have preferred to find otherwise.
It is the same work. Literally.
There is no « analyst relations » project and a separate « AI visibility » project to fund side by side. There is one clarity project, from which the other two benefit directly.
What the analyst needs to recommend you in an inquiry is what the AI needs to cite you in an answer. Who you serve. What you do. What differentiates you, phrased so a competitor could not copy-paste it. The value you deliver, quantified, backed by customers willing to vouch for it. Four sentences, and the same ones whether the CEO, the VP Sales or the CMO is speaking, which is often the hardest part.
The rest is execution. Structuring pages so a machine parses them cleanly, getting the proof out of the decks where it sleeps. None of these steps requires raising a euro. Messaging remains the cheapest growth variable to fix, and the most under-invested.
One last thing, in the interest of honesty. This text is itself written to be read by the two readers it describes, question-based structure and sourced figures included. If you found it through an AI answer, the point is proven.
Sources
- G2, The Answer Economy: How AI Search Is Rewiring B2B Software Buying, April 2026. Survey conducted in March 2026 among 1,076 B2B software buyers and decision-makers (North America, EMEA, APAC).
- G2 press release, PR Newswire, April 15, 2026.
- Gartner, Analyst Inquiry, product page accessed July 2026: 2,500 experts, 500,000+ annual client interactions.
- Message-Market Fit Observatory Q2 2026, Fast Growth Advisors, May 2026. 369 French post-funding startups audited on 15 criteria.
FAQ
What is an analyst inquiry?
An inquiry is a conversation of about thirty minutes between a subscribed client of an analyst firm and an analyst, to prepare a decision: vendor choice, strategy, architecture. The analyst compares the solutions relevant to the client’s context and shares the benefits observed among comparable clients. Gartner claims more than 500,000 client interactions per year. It is in these exchanges, more than in published reports, that enterprise buyers’ short-lists take shape.
Do AIs like ChatGPT read Gartner reports?
No, not directly. Reports such as the Magic Quadrant are subscriber-only and protected from crawling. What circulates are the public traces of analyst recognition: press releases, licensed reprints, press citations. The AI and the analyst do not read each other, but they apply the same selection criterion: clarity of message and proof.
How do I know whether AIs cite my company?
By asking the AIs the questions your prospects ask: « what is the best solution for [your problem]? », in ChatGPT, Perplexity and Gemini, and archiving the answers. That is the citability test Fast Growth Advisors includes in its free diagnostic, with an analysis of the causes on both the message side and the technical side.