I built one of the best sales enablement programmes I have ever seen. I still believe that. The workshops were sharp. The content was relevant. The methodologies were sound. The reps told us the sessions were genuinely useful, and their managers said it was the best thing the company had done for the sales floor.
Then I went and watched the call recordings.
The training had landed in the room. It just hadn't survived contact with the field. Reps who clearly understood the methodology, and could explain it back to you flawlessly, weren't reaching for it on real calls. Not because the training was weak. Because nothing carried it from the workshop into the pipeline.
I brought this to enablement leaders I respected. People who had been doing this for decades. They didn't argue. They said the same three words: "Welcome to enablement."
That was the moment it clicked. We hadn't built a bad programme. We had built a good programme and then handed it to a system that had no way of reinforcing it once the reps walked out of the room.
Three Sources Said the Same Thing This Month
Here is what caught my attention in the last three weeks.
Selling Power published a piece by Justin Zappulla from Janek Performance Group. He named something he calls "the execution gap": the space between what sellers know and what they consistently do when it matters most. He was careful about it too. The point wasn't that training fails. It was that training which lives outside the flow of work "simply does not stand a chance" of being reinforced.
The same week, Gartner published updated predictions. By 2026, sellers will spend 65% fewer hours in training compared to 2022. Not because companies have decided training is worthless. Because event-based delivery, on its own, isn't enough to shift field behaviour. Gartner also predicts AI-driven enablement will deliver 40% faster sales stage velocity by 2029 compared to traditional methods.
And then there is the stat that keeps showing up in every research report I read this quarter: 60 to 90% of sales training is lost within 30 days. Most of what a rep learns this month is gone by the next.
Read that carefully. It does not say the training was bad. It says it was lost. Good content, taught well, with nothing built around it to hold it in place.
I am not cherry-picking these sources. Three independent voices, in the same month, landing on the same thing I have been saying on prospect calls for two years. The training works. What happens after it usually doesn't.
The missing half
Here is the pattern I see.
Everyone in this industry now agrees that training alone does not move field behaviour. Selling Power says it. Gartner says it. The research has said it for years. I have not had a prospect call this quarter where the buyer did not say some version of: "The training is fine. It just doesn't stick."
That second sentence is the one that matters. The training is fine. Often it is excellent. The methodologies are real, the facilitators are good, the reps walk out genuinely sharper. The problem is everything that doesn't happen next.
Look at what the industry keeps building to close that gap. PClub.io, led by the former CRO of Gong, positions around "skill transformation," and builds well-designed, role-specific courses. Yoodli raised $40 million and built a genuinely useful roleplay platform for practising pitches and demos. Proshort maps "contextual AI roleplay" to upcoming deals, which is closer still.
None of that is the enemy. Courses and roleplay are good. A rep should learn the methodology in a workshop and drill it in a simulation before they take it live. That is exactly where training belongs.
But there is still a stretch of road between the simulation and the live call. Between closing the roleplay tab and opening a blank email to a VP you have never spoken to. The training carries the rep up to that line. Then the rep is on their own.
That is the missing half. Not better training. The reinforcement layer that should sit between the training and the field, and almost nowhere has it.
Do the Maths
I want to make this concrete, because the abstract version of "training does not work" lets people nod along without feeling the weight of it.
Take a mid-market sales organisation. 200 reps. Average annual training spend per rep across platforms, content, facilitators, and opportunity cost of time out of the field: conservatively $3,500.
That is $700,000 a year on training.
If 60 to 90% of it is lost inside 30 days, most of that $700,000 buys knowledge that fades before it ever reaches a customer. Not because the spend was wasted on bad training. Because the system let good training drain away.
Now multiply that by every sales organisation in the Fortune 1000. The number gets absurd quickly.
And here is the part worth being precise about. The training itself is mostly excellent. The methodologies work. Gap Selling works. ValueSelling works. MEDDIC works. Run any of them properly and reps get better. The curriculum is not the failure point.
The failure point is architectural. It is the gap between where training happens and where selling happens, and the fact that nothing lives in that gap.
The Real Problem Nobody Wants to Name
Training happens in a training environment. Selling happens in a selling environment. These are two different buildings, and there is no bridge between them.
I call this the Context Gap. And it is the single biggest reason that tens of billions in annual training spend produces so little behaviour change.
The Context Gap works like this. A rep attends a workshop on discovery questioning. Learns a framework. Passes the quiz. Maybe even does a roleplay with a colleague that goes well. They leave the session feeling prepared.
Three days later, they are on a real call with a real prospect who has a real problem they did not anticipate. The framework is somewhere in their notes. The roleplay scenario was nothing like this conversation. The pressure is real. And within 30 seconds, they default to whatever they were doing before the training.
This is not a willpower problem. It is not a motivation problem. It is a physics problem. You cannot learn in one context and execute in another without something bridging the gap.
For years, the instinct has been to make the training itself stickier. Gamify it. Make it mobile, micro, AI-powered. All of that makes good training better, and that is worth doing.
But a better workshop still ends when the workshop ends. The gap is not a content gap, so better content cannot close it. You close it by putting coaching inside the execution environment, so the methodology the rep learned in training is still there, in their hand, on the live call.
What the reinforcement layer looks like
I spent almost three years after that "welcome to enablement" moment trying to answer one question: what would it look like if coaching happened where selling happens, carrying the training forward instead of letting it fade?
Not before the call. Not after the call. Not in a separate platform the rep has to remember to open. Inside the actual workflow. While the rep is writing the email. While they are prepping the deal. While they are processing what just happened on a call.
The model I arrived at, after more failed experiments than I care to admit, is coaching that meets the rep at their real capability on the real deal.
It does not replace the training. It picks up where the training leaves off. Where the rep is still exposed, the AI applies the methodology they were taught at full fidelity and does more of the work, because the muscle memory is still forming. It writes the discovery questions in that methodology. It drafts the email using the framework from the workshop. It structures the deal review the way the curriculum laid out. The rep sees the training they already received, applied to their actual deal, with their actual prospect.
Where the rep is already strong, the AI pulls back. It stops doing and starts coaching. "You missed the gap question here." "This email does not address the problem you identified on the call." "Your next step does not match the deal stage." Same rep, different deals, different levels of support, because the support flexes to the situation rather than running a fixed sequence.
The rep builds capability through guided execution, on top of the foundation the training gave them. The training and the selling stop being two different buildings. There is no context gap because there is no context switch.
This is not theory. I have watched it work across thousands of real coaching interactions. The reps who use it do not just know the methodology. They execute it. On real calls. With real customers. Consistently.
The half nobody is building
What happened this month is genuinely significant. Selling Power, Gartner, and practically every prospect I speak to are converging on the same insight: training builds knowledge, and something else has to convert that knowledge into field behaviour. That is not a coincidence. It is the industry reaching a tipping point.
But there is a difference between naming the gap and building for it. Most of the industry is still pouring its energy into the half that already works. Better workshops. Smarter courses. Sharper roleplay. All worth doing, and none of it the missing piece.
The missing piece is the reinforcement layer. The part that takes the methodology a rep was taught and keeps it alive on the live call, the real email, the deal that just went quiet.
That is coaching that lives where selling lives. It does not compete with training. It finishes the job training starts. That is exactly how AI sales coaching works when it is built properly: inside the workflow, carrying the training the rep already received into the moment it is needed.
So keep the training. Keep your enablement team, your methodology partners, your facilitators. They build something real. The question is what you put between that training and the field, so the work doesn't drain away thirty days later.
Everyone agrees on the gap. Far fewer people are building the layer that closes it.
James Pursey is the CEO of Replicate Labs, an AI sales coaching platform that eliminates the gap between training and execution.