The Room You're Never In: Why B2B Win Rates Are Collapsing

The Stat Nobody's Talking About

A ten-point drop in win rate in twelve months is the kind of number that should be on every GTM leader's dashboard as a red flag. It mostly isn't, because it's easy to explain away.

Budget cuts are real, but budgets were also tight in 2023 and win rates didn't fall this fast. AI disruption is real, but most of the AI activity in B2B marketing right now is about content production, not why deals stall in evaluation. Sales-marketing misalignment is a real and old problem, but it's been used to explain every dip in performance for a decade.

The pattern I keep seeing across engagements is different. It's not that reps are worse or that marketing generates weaker leads. It's that the deal is being decided somewhere the vendor has almost no visibility into, and the pitch that worked in the room doesn't survive the trip outside it.

The Committee Got Bigger, and You Got Smaller

According to Gartner's 2025 research, enterprise B2B buying committees now typically involve 13 to 17 cross-functional stakeholders. Legal, security, finance, the end-user team, procurement, sometimes a second business unit that wants a say. Buyers spend only about 5 to 6% of their total buying-journey time with any single vendor.

Forrester's 2025 Buyers' Journey Survey puts a similar number on it: the average purchase now involves 13 internal stakeholders and 9 external participants, each doing independent research, much of it AI-assisted, on their own schedule.

Do the math on that. If a stakeholder spends 5 to 6% of the journey with you, and there are 13 to 17 stakeholders, you are, at best, one input among dozens, most of whom you will never speak to directly. The deal isn't won in your demo. It's won in the internal debate that happens after your demo, among people who weren't on the call.

This is the first room you're never in: the internal committee conversation where a champion who liked you has to explain, unprompted and unaided, why you're worth choosing over three other vendors who also had a good demo.

The New Silent Stakeholder: AI Chatbots Are Already in the Room

Now add the second room, and this one is new enough that most GTM strategies haven't caught up to it.

6sense's 2025 Buyer Experience Report, published in November and based on roughly 4,000 B2B buyers, found that 94% of buying groups now rank their preferred vendors in order of preference before making contact with a single sales rep. Among the buyers who rank a vendor first on that pre-contact list, 84% go on to purchase from that vendor.

In other words: the shortlist is built, and largely decided, before your team is in the conversation. And a growing share of that shortlist-building now happens inside an AI chatbot, not a search engine.

G2's Buyer Behavior research, part of what it calls the 'Answer Economy' report, found that 51% of B2B software buyers now begin their research with an AI chatbot more often than with Google, up from 29% in April 2025. ChatGPT alone accounts for 63% of that activity. And per Edelman's Trust Barometer, buyers are 3.4 times more likely to trust a vendor that's cited by AI than one they encounter through paid advertising.

Read that last stat again. Trust in an AI-cited vendor beats trust in a paid ad by 3.4x. That's not a minor channel shift. That's a new gatekeeper, and it has no memory of your sales pitch, your charisma, or your best slide. It only has what's publicly written about you, structured clearly enough to be summarized accurately.

Why 'Good Messaging' Fails the Repeat Test

Here's the pattern that connects both rooms, and it's the one I see most often when I'm brought in to fix a stalling pipeline.

A founder or a strong AE can pitch beautifully in person. Tone, timing, a well-placed pause, the right slide at the right moment. It works, in the room, with them presenting.

Then the champion, the person on the buying committee who actually liked you, has to go repeat that pitch to legal, to finance, to a VP who wasn't on the call. They get 60% of it right, half-remember the differentiator, and drop the nuance that made it land the first time. Somewhere in that retelling, the positioning that worked perfectly with you in the room quietly collapses.

That's the champion-retelling problem, and it's been costing B2B deals for years. What's new is that the same test now applies to a machine. An LLM summarizing your company for a buyer has even less context than a distracted champion. It has your website copy, your case studies, whatever's been written about you publicly, maybe a G2 review or two. If your differentiator only exists in your slide deck or your founder's delivery, the chatbot can't find it, and it won't repeat it.

The test in both cases is identical: can your positioning survive being repeated by someone (or something) that wasn't in the room when you explained it well? If the answer is no, you don't have positioning. You have a slide deck that only works when you're presenting it.

Build Positioning for Absence, Not Presence

Most positioning work is built backwards. It's optimized for the pitch, the demo, the moment you have the room's full attention. It should be built for the moment right after that, when you're gone and someone else has to carry it forward.

The framework I use with founders is simple to state and hard to execute: build for absence, not presence. Every piece of positioning has to pass three tests before it leaves the building.

First, the one-sentence test. Can a champion say what you do and why it matters in one sentence, without your help, and get it right? If it takes a paragraph, it won't survive the retelling.

Second, the proof test. Is there one number or outcome attached to that sentence that's specific enough to stick? 'We help companies grow faster' doesn't stick. 'Cut sales cycles by 40% for mid-market fintech' does, because it's concrete enough to be repeated correctly by accident.

Third, the stranger test. If someone who's never spoken to you, human or AI, reads only what's public about your company, do they land on the same one sentence and the same proof point? If your public footprint (website, case studies, reviews, LinkedIn presence) tells a different, vaguer story than your pitch does, that's the gap an LLM will fall into, and the gap a distracted committee member will fall into too.

At one company I worked with, the pitch in the room was strong: a clear technical differentiator explained by a founder who knew exactly how to frame it. But the website talked about the technology, not the business outcome, and case studies were thin on numbers. Champions loved the pitch and then went quiet in the internal debate, because they had nothing concrete to repeat. Once we rebuilt the one sentence, attached a hard number to it, and made sure the public materials matched, the same deals started closing faster, not because the product changed, but because the story survived without us in the room.

Three Moves Israeli Founders Should Make This Quarter

None of this requires a rebrand or a bigger content team. It requires discipline about what has to survive your absence.

• Run the retelling test. Ask your best champion in an active deal to explain your value in one sentence, cold, without prep. Whatever they say back is your real positioning, not the one on your slides.

• Audit your AI footprint now. Ask ChatGPT and Perplexity what your company does and who it's for. If the answer is vague, generic, or wrong, that's what a growing share of buyers are reading before they ever contact you.

• Write for the 5% room, not the deck. Your website, case studies, and one-pagers need to carry the full argument on their own, because 94% of buying groups are building a shortlist before a human from your team says a word.