I have bought sales training. Twice, with my own budget, for teams I was accountable for. Both times the room was loud and grateful on Friday afternoon, both times somebody called it the best off-site we'd ever run, and both times I opened the pipeline in March and found the same deals stalled at the same stage for the same reasons. Nobody was lying about the value. The training simply did not survive contact with a Tuesday.
Twenty years of running commercial organizations later, I think I know why, and it isn't that the material was bad. Most sales training content is fine — the frameworks have been sound since the 1980s. It fails because of its shape: an event, delivered to individuals, about hypothetical deals, with nobody on the other side of it owning what happens next. Selling is not knowledge. Selling is habit under pressure, and you don't change a habit with a slide deck any more than you get in shape by attending a lecture about squats.
Here is what I have actually watched survive.
Run it on live pipeline, or don't run it
The strongest predictor of whether training sticks is whether the practice deals were real. Role-play with an invented prospect teaches a rep to perform for the room. Working their own stalled opportunity — the one with a decision-maker they have never met and a champion who has stopped replying — teaches them the thing they will need on Thursday.
So we build the sessions around the CRM. Every rep brings three open opportunities: one they think will close, one they think is dead, and one they cannot read. By the end of the day, each of those has a written next step, and the next step is not "follow up." It is a named person, a specific question, and a date. The measurable output of good training is not enthusiasm. It is a pipeline that moved while everyone was sitting there.
Record the calls, then review them line by line
Every rep believes they run a good discovery call. Almost none of them have heard one of their own calls end to end. The first time a rep listens to themselves talk for sixty percent of a discovery conversation, answer a pricing question that was never asked, and skip past a buying signal in the customer's third sentence, the argument is over. I have never had to persuade anyone after that; the tape does the persuading.
The practice is unglamorous: record everything, pick two calls a week per rep, review them together with the transcript open, and stop at the specific moment the call turned. Not "be more consultative." There. At 14:20. She told you the budget cycle closes in November and you kept talking about integrations. Coaching that names a moment changes behavior. Coaching that names a virtue does not.
Train the managers separately, and train them first
This is the one most companies skip, and it is the reason the rest evaporates. A sales manager who did not attend, or attended as a spectator, will go back to asking "how's it looking?" in the Monday meeting — and reps optimize for the question they are asked. If the question never changes, the behavior never changes, no matter what happened at the off-site.
Managers need their own material: how to run a pipeline review that inspects the deal rather than the forecast, how to coach a call without taking it over, what to do with the rep who is busy and not productive. We run that session apart from the team, and we run it first, so the manager walks into training already knowing what they will be reinforcing for the next quarter. A program that reaches only reps stops working the month after we leave.
Write the playbook down while it is hot
By Friday afternoon a team has generated an enormous amount of value in the room: the objection that finally got handled well, the discovery question that opened three accounts, the way one rep explains pricing that nobody else had thought to copy. Almost all of it is lost by the following quarter unless someone writes it down that week.
The written playbook is the difference between training a team and training a company. Messaging, qualification criteria, the objections in your market and the language that works on them, the sequence that follows a first meeting — in one document your next hire can run from in week one. Then the CRM and the follow-up sequences get configured to match it, so the playbook is not a PDF in a drive folder. It is the path of least resistance inside the tool the rep already has open.
Put AI on the part of the job reps hate
Here is where the last two years genuinely changed the math, and where our sales training and our AI training stopped being two separate programs.
Ask any quota-carrying rep where their week goes and you will hear the same list: account research before the call, notes after the call, the CRM fields nobody fills in, the follow-up email that gets written three days late, the proposal assembled from the last proposal. None of that is selling. All of it is necessary. It is also, almost precisely, the work a well-configured assistant does well — research a company and its people into a one-page call plan, turn a recorded conversation into structured CRM notes, draft the follow-up in the rep's own voice within the hour, and assemble the first pass of a proposal from your actual pricing.
Two cautions, learned the hard way. First, every one of those workflows needs a human checkpoint before anything reaches a customer — one fabricated number in a proposal costs far more than the hour it saved. Second, do not hand a team a chat window and call it enablement. The teams that get real hours back are the ones where each role got a purpose-built assistant, trained on their own accounts, in a session where they did the work themselves. Everyone else has a subscription and a story about how they tried it once.
Done properly, this is the rare intervention that gives a rep back four or five hours a week — and reps who have four hours back make more calls. That is the entire mechanism. There is nothing clever about it.
What to look at in ninety days
Enthusiasm is not a metric, and neither is training attendance. Ninety days out, I look at four things: the number of first meetings booked per rep per week, the percentage of opportunities with a written, dated next step, stage-to-stage conversion on the two stages you targeted, and average time from first meeting to proposal. If none of those moved, the program was entertainment, whoever delivered it.
You should hold anyone you hire for this — including us — to exactly that standard, in writing, before the first session.
The uncomfortable part
The reason training fails is rarely the trainer. It is that companies buy it as a substitute for a decision they have been avoiding: the comp plan that rewards activity over outcomes, the manager who was promoted for selling and never taught to coach, the pricing nobody can defend, the two reps everyone knows are not going to make it. A workshop cannot fix any of that, and a good one will surface all of it by lunch on the first day.
That is the real test of whether training is worth buying. If the person running it hands you a certificate, you bought an event. If they hand you an uncomfortable list of things about your own commercial organization — and then stay close enough to help you work through it — you bought something that will still be paying you in March.