The people who research you are not the people who sign

The audience that reads your material most carefully has no signing authority. And these days it does not read the material itself.

A target group that was parked on purpose

In a project meeting recently — a provider of leadership development in the DACH region, mid-sized, four defined buyer groups — an obvious expansion was on the table. Two partners had pushed for it: address the specialist and development leads one level below the executive floor. The search had already found a few hundred reachable contacts. The team parked it anyway, with a one-sentence rationale: this role researches, it does not sign. The estimate in the room was that maybe one in ten of those contacts holds signing authority at all.

That decision reads as conservative until you put it next to the numbers from the same week, which appear to say the exact opposite.

The numbers that seem to contradict it

A survey of roughly 18,000 business buyers found that 94% used AI during their most recent purchase, 55% use it to compare vendors against each other, and 47% build the internal business case before contacting any vendor at all. A separate analysis puts 80% of B2B technology buyers as using AI agents for buying tasks — worth noting that the underlying fieldwork is from the previous year, even though the write-up is new. And the median buying committee for deals above 50,000 dollars has grown to 11.2 people.

Read quickly, that says: the research phase is where the deal is decided, so target the researchers. Read carefully, it says something narrower and more useful.

What actually changed

The person one level below the signature used to be the bottleneck and the filter. They read the material, summarised it, and carried a recommendation into the room. That is the job that changed. They no longer summarise what a provider can do — an assistant does, from whatever it can retrieve.

So the target-group question splits into two questions that people keep answering as one:

• Who decides? Unchanged. Still the same small set of roles, now surrounded by a committee of eleven.
• What gets retrieved? Completely changed. Whatever a machine can find, parse, and quote — on the open web, without a login, without a form.

Buying a list of researchers optimises for the first question using the mechanics of the second. That is a category error, and it is expensive: you pay per contact for a role whose influence now runs through material you have not published.

The form gate is the part that hurts

The uncomfortable consequence lands on the oldest reliable mechanic in B2B marketing. What sits behind a form does not exist for a retrieval agent. It cannot be read, so it cannot be cited, so it does not appear in the business case that gets built before anyone contacts you.

A gate that used to generate leads now makes you invisible in the phase where the shortlist is written. That is not an argument against lead capture in general — it is an argument about sequence. Capture after the citation, not instead of it.

Two readers, two kinds of text

The practical answer is unglamorous and mostly editorial work.

• For the retrieval path: short, self-contained blocks that answer one buyer question each. Numbers with a source and a collection period. Named trade-offs, including where you are the wrong choice. Prices or at least price logic. These are written to be quoted out of context, because that is exactly what will happen to them.
• For the signature: decision material. What it costs, what it replaces, what happens in the first ninety days, who carries the risk. Nobody quotes this. One person reads it, once, and has to be able to defend it to ten colleagues.

One generic capability asset serves neither. And the team that parked the obvious target group was not being cautious — they had simply separated the two questions correctly, which is more than most content plans manage.