AI-Informed Buyers: Why 69% Still Check With a Rep

Quick answer: B2B buyers now show up to the first call with an answer already formed, assembled from AI tools before a seller was involved, and what they want from that seller is a check on it. Gartner’s survey, announced May 2026, found 69% of B2B buyers turn to sales reps to validate AI-generated insights. That reprices the job. A rep whose only skill is presenting information is now worth less than the free tool the buyer used the night before, while a rep who can look at a buyer’s conclusion, find where it’s wrong for their business, and say so credibly is worth more than they were a year ago. That check is Problem-Centric Selling, A Sales Growth Company’s approach of working the buyer’s problem rather than your product. Correcting a smart buyer’s wrong answer takes business acumen, and most orgs never taught it, never hired for it, and don’t coach it. That makes this a hiring decision, a training decision and a coaching standard, which puts it on the CRO’s desk and nowhere else.

Key takeaways

• Gartner’s May 2026 survey found 69% of B2B buyers go to sales reps to validate AI-generated insights. Buyers aren’t skipping sellers, they’re using them differently.
• The information part of the sale moved in front of the first call. Whatever your reps were going to explain, the buyer already had a version of it.
• The value now sits in the check, not the pitch: is this answer right for this company, this situation, these constraints?
• Telling a confident buyer their conclusion is wrong takes business acumen. That is a hiring bar most scorecards don’t test.
• Discovery gets harder, not easier. A buyer with a pre-formed answer will steer the conversation toward confirming it.
• This is a system decision. Who you hire, what you train, what managers coach to. Not a prompt library.

How is AI changing B2B buying behavior?

AI moved the information part of the sale in front of the first call. Buyers research vendors, build comparisons, draft requirements and reach a working conclusion before anyone in sales knows the account exists. That work used to be the seller’s opening thirty minutes, and it was the reason a lot of first meetings got booked at all.

We’ve all watched some version of this happen. The buyer joins the call already fluent, already holding a shortlist, already using language that used to signal a late-stage deal. Then they ask a question that has nothing to do with what your product does and everything to do with whether the thing they already decided is actually right.

Forrester’s 2026 predictions put trust at the center of B2B for exactly this reason: when everyone can generate a credible-sounding answer in seconds, credibility itself becomes the scarce thing. The buyer has more information than ever and less certainty about which parts of it apply to them.

Do buyers still need sales reps if they use AI?

Yes, and the data says so plainly. Gartner’s survey, announced May 2026, found that 69% of B2B buyers turn to sales reps to validate AI-generated insights, and Demand Gen Report’s coverage of that survey framed it as human sellers closing the confidence gap. The buyer isn’t coming to you to learn. They’re coming to you to find out whether what they learned holds up.

McKinsey has made a related argument, that AI expands what commercial teams can do rather than replacing them. True, but soft. The sharper version is this: AI replaced one specific rep skill, the ability to explain things, and left the harder ones alone. If a seller’s whole contribution was knowing the product better than the buyer did, that gap just closed, permanently and for free.

It’s a bit like a patient walking in with a diagnosis they got online. They don’t want it read back to them. They want someone who has seen two hundred cases to tell them which part of it is right, which part is wrong, and what it means for them specifically.

What should sellers do when a buyer arrives with AI research?

Check it out loud, against evidence from the buyer’s own business. That means getting the buyer’s conclusion on the table early, understanding what it’s built on, and then testing whether the assumptions underneath it are true here.

Three things have to happen in that conversation, and none of them are a pitch.

• Get the answer stated. Ask what they’ve concluded and why, before you present anything. Otherwise you’re arguing with a position you can’t see.
• Find the assumptions. Every AI-generated recommendation rests on inputs: a benchmark, a category definition, a set of typical costs, an assumption about how a team operates. Those are generic by construction. Yours is a specific company.
• Test them against the buyer’s own numbers. Not against your case studies. Against what is actually happening in their business, which is the only evidence that can move a decision they already made.

This is ordinary problem-centric work, done under new conditions. Problem-Centric Selling was always about the buyer’s problem rather than your product, and the reason it holds up here is that AI is very good at describing categories of problems and very bad at knowing whether one exists inside a particular company. The evidence gap is where the seller lives now. If your reps can’t produce buyer input data, facts from the buyer’s own operation, the conversation stays theoretical and the AI answer wins by default.

What happens when the buyer’s AI answer is wrong?

Somebody has to tell them, and that is the skill this shift is actually pricing. A wrong answer delivered confidently by a tool the buyer trusts is harder to move than no answer at all, because the buyer has already invested reasoning in it and, often, already shared it internally.

Correcting it requires knowing how businesses work. How the cost the buyer identified actually shows up on their P&L. Why the fix they picked fails in a company with their structure. What the second-order effect is on the team next door. That’s business acumen, and it isn’t product knowledge, and it isn’t charisma, and you cannot fake it in front of someone who has spent the last two weeks reading.

Here’s the part that should worry a revenue leader: a rep who lacks it doesn’t argue and lose. A rep who lacks it agrees. They nod along with a flawed conclusion because they have nothing credible to counter it with, and the deal proceeds toward a solution that won’t produce the result, which is how you end up with a closed-won that churns in year two.

What does this change about who you hire?

It changes what a strong candidate looks like, and most hiring scorecards haven’t caught up. The traditional profile rewards activity, presence, tenure in the category and a rolodex. None of those tell you whether a person can look at a buyer’s reasoning and find the flaw in it.

What you’re hiring for now is judgment under someone else’s confidence. Can this person hold a position against a buyer who has done real homework, without being either a pushover or a jerk about it? Can they reason about a business they don’t work in?

That’s testable in an interview if you bother to test it. Give a candidate a real scenario with a plausible but wrong conclusion baked into it and watch what they do with it. The ones who go straight to the product are telling you exactly what they’d do on the call. If your process doesn’t have a stage that does this, the fix is a better scorecard, and we publish a candidate hiring scorecard you can adapt.

How do you train and coach reps for AI-informed buyers?

Train the acumen and coach the conversation, in that order, and stop treating either one as a one-time event. Product training does nothing here. A prompt library does nothing here. The skill is reasoning about a buyer’s business well enough to disagree with them usefully, and that gets built through repetition against real deals with a manager who knows what good looks like.

Two pieces have to exist for this to work. The first is a defined standard for the skill itself, which is what the Skills layer of a sales system is for: what competent looks like, written down, so a manager and a rep are grading against the same thing. The second is a cycle that catches the rep’s actual behavior and corrects it, which is what improvement loops do. Without the loop, training is an event people remember for six weeks.

What managers should be listening for in call reviews is narrow and specific. Did the rep surface the buyer’s existing conclusion before presenting anything? Did they identify what that conclusion assumed? Did they bring the buyer’s own numbers to the test, or did they bring a case study? A manager who only asks where the deal is and when it closes will never see any of this.

What does it cost you if you ignore this?

It costs you the first call, and then it costs you the deal, in ways that never show up as a clean loss reason. Here’s the bill, one line at a time.

Reps who can only recite what the product does. – The buyer read a better version of that before the meeting, for free.

Discovery that stops at “what are you looking for.” – You learn the buyer’s AI-generated conclusion and nothing underneath it.

Nobody testing the buyer’s assumptions. – You compete on the buyer’s terms, which means price.

Managers inspecting the forecast and calling it coaching. – The one skill the job now requires never develops.

Hiring for territory relationships and tenure. – You’re buying the skill the market just repriced downward.

Agreeing with a wrong conclusion to keep the deal moving. – A closed-won that doesn’t produce the result, and a renewal conversation you lose.

None of those read as an AI problem on a loss report. They read as price, or timing, or no decision.

Where should you start?

Find out how your buyers actually want this handled before you redesign anything. ASG surveyed more than 1,200 B2B buyers on what they want from sellers, and the report, How Buyers Want to Be Sold, is free at salesgrowth.com. Then look at your own hiring stage and your own call reviews and ask whether either of them tests the skill this market now pays for.

Frequently asked questions

Does this mean companies need fewer sales reps?

It means they need different ones. The work that’s disappearing is information delivery, and the work that’s growing is judgment applied to a specific buyer’s situation. An org staffed entirely with presenters will need fewer people. An org staffed with sellers who can test a buyer’s reasoning will find those people busier than before.

Should reps be using AI themselves?

Yes, for the same reason buyers do, to prepare faster and cover more ground before a call. It doesn’t change what happens in the meeting. The seller’s contribution is still the part AI can’t produce: evidence from inside the buyer’s business and a judgment about what it means.

How do you know if a buyer has done AI research if they don’t mention it?

Ask what they’ve already looked at and what they concluded. Most buyers will tell you. The tell is a buyer who arrives with unusually precise category language and a pre-built comparison, then asks confirmation questions rather than open ones.

Does this change the discovery process?

It adds a step at the front. Before working the buyer’s problem, surface the conclusion they’ve already reached, because a buyer with a formed answer will steer discovery toward confirming it and you’ll mistake their agreement for alignment.

Is business acumen something you can actually train?

Yes, though not in a workshop. It’s built by repeatedly reasoning about real businesses, real financials and real operating constraints, with feedback from someone who can tell the difference between a good read and a confident guess. That’s a coaching cadence, not a curriculum.

Where does this leave marketing?

In a similar position. If AI tools are summarizing your category before a buyer talks to anyone, the content that matters is the content that gives a buyer something they can’t generate: specifics, evidence and honest limits.

About the source

This article is published by A Sales Growth Company (ASG), the creator of Problem Centric® Selling and the architect of the Problem-Centric Operating System (PCOS™). Keenan is the CEO of ASG and the author of Gap Selling.

Sources

• Gartner survey finding that 69% of B2B buyers turn to sales reps to validate AI-generated insights, announced via Businesswire, May 20, 2026.
• Demand Gen Report coverage of the same Gartner survey, May 26, 2026, “Gartner: AI Is Reshaping B2B Buying, but Human Sellers Still Close the Confidence Gap.”
• McKinsey on the future of B2B sales with AI, arguing AI expands rather than replaces commercial teams.
• Forrester, Predictions 2026: Trust Gets Tested For B2B Marketing, Sales, And Product Leaders.
• Problem-Centric Selling, Buyer Input Data, the Skills layer and improvement loops are ASG frameworks.