Ask a sales team to walk you through their process and they can do it without hesitation. First call, discovery, proposal, close. Ask them to walk you through the buyer’s process during that same stretch of time, and the room gets quiet.
That gap is worth sitting with, because it’s rarely a knowledge problem. It’s rarely even a design intention problem. Most sales leaders genuinely try to build stages around what the buyer does. “Proposal signed” instead of “proposal reviewed.” “Contract executed” instead of “contract under legal review.” The instinct to mark stages by buyer milestones, not rep activity, is usually there.
What gets lost is the actions in between. Sending a proposal is something a rep does, easy to log and timestamp. The buyer signing it is the milestone, also easy to log. What’s missing is any sense of how long that gap should reasonably take, given what the rep knows about how this buyer will make the decision. Three weeks between sending and signing might be completely normal for one customer. Yet a real warning sign for another. Without a sense of what’s normal for each one, both look identical sitting in the pipeline.
Time Gets Logged. Pace Doesn’t.
Picture a typical pipeline review. Every deal has a stage, and most of those stages really are named after something the buyer did, not something the rep did. On the surface, that looks like a process built around buyer behavior, and in a narrow sense, it is. But ask how long a deal has been sitting between two milestones, and whether that length is normal for this specific customer, and the answer usually gets vague fast.
That’s because the system tracks when things happened, not whether the pace is right for this buyer. A large, cautious customer with several layers of internal sign-off might reasonably take two months between a proposal going out and coming back signed. A smaller, faster-moving customer might take two weeks. Judge both against the same generic benchmark, or against no benchmark at all, and neither one tells a rep anything useful about whether the deal is actually on track.
I’ve watched teams get genuinely disciplined about naming their stages after buyer milestones and still lose deals they should have won, because naming the milestone isn’t the same as having a reasonable expectation of how the customer sitting in front of it is likely to move. Everything that happens in between still relies on a rep’s gut read, and gut reads are inconsistent by nature.
This is the quiet cost of a system built around buyer milestones without a sense of buyer-specific pace. It doesn’t fail loudly. It just treats every deal as if it’s moving at the same speed, right up until the one that wasn’t.
What AI reveals about the gap
Here’s where it gets interesting, because AI doesn’t create this problem. It just makes it visible faster.
Run a pipeline through an AI summary tool and it will hand back a clean, confident number. No hedging, no uncertainty in the tone, just a tidy read of what’s in the system. And that confidence does something to a room. People who’d been quietly unsure about the forecast tend to relax a little, not because anything was actually verified, but because it sounded verified.
AI can only work with the signal it’s given. If a stage definition is stretched for months to make a deal look further along than it is, AI won’t catch that. It will simply report it back more smoothly than your team ever did, because clean reporting is what the tool is built to do.
I’ve started thinking about AI less as a second opinion and more as a mirror. It doesn’t tell you the truth about your pipeline. It tells you, very clearly, what your pipeline already believes. If nothing in the system ever captured what a normal pace looks like for this buyer, AI won’t invent that judgment either. It just delivers the same blind spot with more polish and less doubt than it deserves.
The risk was never that AI would get the forecast wrong. It’s that a confident-sounding summary can quiet the doubt a leader should still be sitting with.
Where the real decisions happen
This is where the gap turns into a leadership problem rather than a data problem. A deal sitting quietly for three weeks isn’t automatically a red flag. It might be nothing. Whether it’s nothing depends on what the rep knows about how this buyer is making the decision. Three weeks might be expected, or it might be the first sign something has changed. That judgment call rarely gets made out loud. It gets made by feel, deal by deal, rep by rep.
So a deal is still predicted to close because the stage moved, not because anyone stopped to confirm that the buyer is still moving as expected. A stalled account gets pushed to next week because pushing it feels like patience, not because anyone stopped to ask whether the delay still makes sense given what they know about the buyer’s decision process. None of these moments look consequential on their own. They feel like reasonable calls made with incomplete information, because they are.
But they compound. By the time a leadership team sits down to talk about the number, most of those calls have already been made quietly, without anyone testing them against what’s actually normal for the customer in question. The system didn’t drift because someone made a bad decision. It drifted because the information needed to make a good one, a sense of this buyer’s pace against their own history, was never something anyone was collecting.
Building a Pipeline That Tracks Both
None of this gets fixed by adding more milestones or a smarter AI summary. Both just add more precise timestamps to a system that still has no sense of what normal looks like for each buyer.
What actually closes the gap is developing a reasonable expectation, deal by deal, of how the buyer’s decision is likely to unfold, and treating a meaningful deviation from that pace as a signal worth examining rather than a delay worth waiting out.
That’s a harder habit to build than a new stage definition. It asks reps and leaders to stop judging every deal against a single average and start asking whether the pace still makes sense for the buyer actually sitting across the table.
The question becomes whether your pipeline can tell the difference between a deal that’s slow and a deal that’s stalling.
