Ask what the AI strategy will decide, who owns it and when it lands, and whether any of that decides your purchase. Separate the two, become an input with a small dated pilot, or park the deal. The library's own move.
"We're figuring out our AI strategy first." The buyer liked the product. Now everything waits for a strategy with no owner you know of and no date.
Ask what the strategy has to decide, who owns it and when it is due. Then check whether your purchase depends on any of those decisions. If it does not, say so, and separate the two: the problem you solve exists whatever the strategy says. If it does, offer to be an input to the strategy rather than something it rules on later, for example with a small pilot, fixed in scope, with a date on which both sides decide go or no-go. If nobody owns the strategy and there is no date, treat the answer as no for now, and park the deal with a trigger, as in Park with a trigger.
No seller publishes this move. It is the library's own, put together for a stall that research on buyers has only recently started to measure.
The questions, in order. What will the AI strategy decide? Who owns it, and when does it land? Which of those decisions would change whether this purchase makes sense? What would the strategy group need to see from a small, dated pilot? The answers separate a real dependency from a way of saying not now.
Nobody has tested the move. No study compares sellers who asked these questions with sellers who waited for the strategy.
The stall is real and growing. Gong's trend data, from 33.5 million deals at 3,398 companies since February 2024, shows deals where AI comes up at all up by 85%, deals where a buyer mentions cutting a budget they had already disclosed up by 41%, and deals where someone worries about AI replacing staff up by 237%. Gong sells the software that tracks these mentions, and gives only the rise.
Buyers have reason to be cautious. S&P Global Market Intelligence surveyed 1,006 IT and business professionals in North America and Europe and reported in May 2025 that the share of companies abandoning most of their AI projects before production had risen from 17% to 42% in a year, and that on average 46% of projects were dropped between proof of concept and wide use. McKinsey's survey of 1,719 people in 97 countries, published in August 2026, found the share reporting any profit impact from AI unchanged from a year earlier at 37%, and about one in five saying the running costs of AI had limited their use of it. S&P sells market research, and McKinsey sells AI consulting.
People wait for information that would not change their choice. Anthony Bastardi and Eldar Shafir ran a series of experiments, published in 1998, in which people chose to wait for information that looked relevant but would not have changed their decision if they had had it from the start. Once they had waited for it, they used it, and made choices they would not otherwise have made. An AI strategy can be exactly this kind of information for a purchase that does not depend on it.
A hard choice makes people put it off. Amos Tversky and Eldar Shafir's experiments on deferral, described under When the buyer keeps asking for more, ask what it is for, found that people facing a difficult choice chose to wait and look for more options.
The market figures describe the climate, not what a seller's questions do to a deal, and they come from firms that sell research, consulting or call-analysis software. The waiting experiments used hypothetical choices, not purchases. Nobody has measured this move.
Ask what the AI strategy will decide, who owns it and when it lands, and whether any of that decides your purchase. If nothing does, separate the two. If something does, become an input with a small, dated pilot. If there is no owner and no date, park the deal with a trigger. The research says people often wait for information that would not change their choice, so find out which kind of wait this is.