Gartner published three findings inside about twelve months that are usually quoted separately, and quoted separately each one supports a different strategy.

In March 2026 it reported that 67 percent of business buyers say they prefer a rep-free buying experience. In May 2026 it reported that 69 percent of business buyers turn to sales representatives to validate insights generated by AI. In August 2025 it had predicted that by 2030, 75 percent of business buyers will prefer sales experiences that prioritise human interaction over AI.

The reading that resolves them

Taken as a set rather than a contradiction, they describe a buyer who wants to do the work alone and wants a person present at the moment the work carries risk. Those are not competing preferences. They are the same preference expressed at two different stages of the same decision.

The rep-free finding is about the research, comparison and shortlisting phase, where a person adds friction and little else. The validation finding is about the moment the buyer has an answer and has to stake something on it being right. The 2030 prediction is a statement that as more of the first phase is automated, the second phase matters more rather than less.

Why firms get this wrong in both directions

A firm that reads only the first finding removes people from the process, replaces them with self-service and automation, and finds that deals stall at exactly the point where a buyer needs a human to take responsibility for an answer.

A firm that reads only the third finding concludes that the human relationship still wins and leaves its early-stage presence untouched. It then loses to competitors at the shortlist stage, where no human was ever going to be involved, and never learns why because the deals it loses there do not appear in its pipeline.

What we take from it

This is the closest thing to external support for the position this firm was already built on, which is that the modern buyer is self-educated and still relies on human validation at the moments of real risk, and that both halves have to be designed for.

In practice it argues for a specific allocation. The research and comparison phase should be servable without a conversation, which means the answers a buyer needs have to exist in public and be findable. The validation phase should be staffed by someone able to take a position and be accountable for it, which means it should not be the cheapest available seat.

It also argues against the middle option most firms run, which is a process that is neither self-serviceable nor properly staffed: a gated document, a form, a qualification call, and a person who arrives without a view.