Sixteen Interactions. The Number That Refuses to Move.

The number was on a slide. Sixteen.

The VP of Sales looked at it the way you look at a line item you plan to cut. Sixteen interactions per buyer with the vendor who wins the deal. His read was immediate and confident: too many. Two years into building a leaner, self-serve motion, sixteen touches felt like the old world hanging on. Something to drive down. A cost.

He read the number exactly backwards.

Sixteen is not the number of touches the buyer tolerates before they buy. Sixteen is the number of touches the buyer has with the vendor they choose. It is the winner's number. And it has barely moved in three years, through the single biggest change in how buyers research anything since the search engine.

The number that refuses to move

Buyers now average sixteen interactions per person with the winning vendor. That figure is down one from 2024 and statistically identical to 2023, according to 6sense's 2025 B2B Buyer Experience Report. Three years. Same number.

Now sit that number next to what else changed in the same window. In that same report, 94% of buyers said they used AI in their buying process. Nearly every buyer added a machine to their research. And the number of times they needed to touch the winning vendor did not move.

Read that sequence slowly, because the headlines skipped it. AI arrived. Buyers adopted it almost universally. The interaction count held. Whatever AI changed, it did not change how many times the buyer has to engage the vendor they eventually trust with the decision.

AI changed what gets researched inside each interaction. It did not change how many interactions the winner earns.

The narrative you built your org around

Here is the part where I tell you the narrative you were following was not stupid. Because it was not.

For a decade, the best research in the field pointed one direction. Buyers spend only 17% of the total buying process with any single vendor, and roughly 80% of it self-directed, Gartner reported. The buyer was doing the work alone, in the dark, long before a rep got a meeting. Then the preference data caught up to the behavior. As of June 2025, 61% of B2B buyers said they preferred a rep-free experience, and the number was climbing.

So you did the rational thing. You built for it. You trimmed the SDR touches that felt intrusive. You poured budget into self-serve content and product-led motions. You braced for AI to shrink the funnel further, and you staffed as if the buyer of 2026 would want you in the room as little as possible.

That was not a bad bet on the evidence you had. It was the responsible read of a real trend. The problem is not that you believed it. The problem is what happened when the trend met the touch count.

Where the narrative breaks

The rep-free story made one quiet substitution, and everything downstream inherited it. It treated "the buyer self-directs their research" as if it meant "the buyer decides alone."

Those are not the same sentence.

A buyer can run 60% of the buying process through digital research, form opinions before you know they exist, walk in already leaning toward a vendor, and still need sixteen interactions with that vendor to actually commit. Self-directed research and heavy vendor engagement are not opposites. They are the same buyer, doing both.

The 2025 data makes the substitution impossible to ignore. In the same year buyers went all-in on AI, 62% of them said they needed sellers to clarify AI capabilities, and 58% engaged vendors earlier than usual specifically to get their AI questions answered (6sense, 2025). Think about what that does to the rep-free thesis. AI was supposed to be the thing that removed the seller. Instead, AI became a new category of question the buyer could not resolve on their own, so they came to the seller sooner. The machine that was going to replace you generated demand for you.

The buyer added AI to the front of the process. The sixteen interactions stayed right where they were.

Why the number won't move

If you want to understand why sixteen refuses to drop, you have to give up one assumption: that a sales interaction is a delivery mechanism for information.

That was never what it was.

The most useful reframing of the seller's job in years comes from a 2025 paper in the Journal of Marketing. Kalwey and colleagues redefine the salesperson's core role plainly: to coordinate every buyer-seller touchpoint in a way that builds the buyer's decision-making competence. Not to drive the interaction. Not to close. To make the buyer more capable of deciding.

Hold that against a large language model. AI is extraordinary at supplying information. It is a research accelerant, a summarizer, a first-draft shortlist generator. What it cannot do is sit across from a buying group and build the shared confidence a five-person committee needs to say yes together. It cannot read the room where the CFO and the RevOps lead disagree and reframe the problem so both can move. It cannot absorb the specific fear the buyer has about implementation and answer it in a way that sticks.

That work takes interactions. It takes more than one, because confidence is built, not transferred. That is why the number is sixteen and not six, and it is why sixteen held through the AI wave. The interactions were never the cost of the sale. They were the sale.

You don't need fewer sellers. You need different ones

So the strategic error was never "we invested in self-serve." Self-serve is real and the buyer wants it for the parts of the buying process that are genuinely self-servable. The error was concluding that a buyer who researches alone wants to buy alone, and then cutting the human capacity that builds the decision.

The 2026 buyer does not need fewer sellers. The 2026 buyer needs a different seller.

The old seller was an information source, and the buyer has better information sources now. The buyer will route around that seller every time, and should. The seller who survives is the one who does the thing the buyer cannot get from a chatbot or a comparison grid: clarify what the AI got wrong, frame the criteria the buyer has not thought to weigh, and hold the buying group together through the messy middle where deals actually die. That seller is not a cost to minimize. That seller is why one vendor got sixteen interactions and the other got left on read.

This is the whole premise of measuring buyer readiness instead of counting activities. The sixteen interactions are where the four dimensions of buyer readiness get built or left broken. Cut the interactions and you have not saved money. You have removed the only place readiness gets made.

The firm that named rep-free is already calling the reversal

Here is the detail that should settle it. The pendulum is not just failing to swing toward rep-free as fast as predicted. The analyst that named the trend is now forecasting its reversal.

In August 2025, Gartner predicted that by 2030, 75% of B2B buyers will prefer sales experiences that prioritize human interaction over AI, as AI fatigue sets in on complex, high-stakes purchases. The same firm whose 61% rep-free stat you built your restructuring around is telling you the preference inverts inside five years, precisely on the deals that matter most.

That is not a reason to ignore the rep-free preference. It is a reason to stop treating it as a destination and start treating it as one reading on a moving dial. The buyer wants to self-serve the parts that are simple and wants a capable human on the parts that are not. AI raised the stakes on getting the hard part right. It did not remove the hard part.

What to stop doing on Monday

If you have been planning for the rep-free future, three things are worth stopping this week, before the next planning cycle bakes the misread in deeper.

Stop measuring interactions as a cost to drive down. The question is not "how do we reduce touches." The question is "does each touch build the buyer's ability to decide, or does it just consume their time." A team chasing fewer interactions will happily cut the three touches that would have won the deal. Count readiness built, not meetings avoided.

Stop staffing for a buyer who self-serves to the finish line. That buyer is rare on complex deals, and the data says they are not getting more common at the touch-count level. Staff for the seller who earns the sixteenth interaction: the one who clarifies the AI questions, frames the evaluation, and works the buying group. That is a different hire and a different enablement program than the one a rep-free plan builds.

Stop reading "the buyer used AI" as "the buyer needs you less." AI has reshaped the buying committee and the research that feeds it, and it has raised the number of questions in the room. More questions, not fewer, is what generates the sixteenth interaction. The buyer using AI is often the buyer about to call you earlier.

The number, reread

Go back to the slide. Sixteen.

The VP saw a cost. What the number actually shows is a mechanism. Sixteen is how many times the buyer engaged the vendor whose people made them ready to decide. The vendors on the other side of that deal, the ones who got six interactions and a polite no, did not lose because they had too many touches. They lost because they had too few of the right ones, or because they read their own sixteen the way the VP read the slide, as friction to remove.

The number refuses to move because the thing it measures never went away. The buyer was never trying to avoid you. The buyer was trying to get ready. Sixteen is what getting ready costs, and the vendor who understands that stops trying to shrink the number and starts trying to earn it.

That is the whole shift. Not fewer interactions. Better ones. The number is not the problem. The number is the scoreboard.

See where your deals are actually stalling

Most pipelines are full of deals with plenty of activity and no readiness underneath it. The buyer readiness assessment takes four minutes and shows you which of the four dimensions your deals are stalling on, so you can tell the difference between an interaction that built the decision and one that just filled the calendar. Start with the deal that went quiet after the demo.