Why are sales managers coaching less in 2026?

Top sales managers are coaching 45% less than they were three years ago, according to Mindtickle's fourth annual State of Agentic Revenue Enablement report (Mindtickle, 2026), which draws on platform activity from 400+ companies, 1.5 million users, and more than a million call recordings. The report's read is that AI copilots absorbed the load, freeing managers for higher-value work.

Key Takeaways

  • Top managers are coaching 45% less than three years ago (Mindtickle, 2026), and AI reviewed 465,000 role-play submissions in 2025, work that would take one person 16 months full-time.
  • Mindtickle frames the drop as capacity freed up. We think it's a coaching habit that was already gone, now visible for the first time in someone's own data.
  • Reviewing a role-play submission against a rubric is not the same act as coaching a rep, and the report's own numbers show which one managers actually stopped doing.

I have read that stat three times now and it still lands the same way: that is not a story about AI freeing anyone up. That is a story about a manager's calendar finally telling the truth.

A number that confirms what every sales floor already knew

Nobody who has run a sales team will read "coaching down 45%" and feel surprised. I ran a team where one manager had eleven reps, and by the third quarter his Friday coaching block had quietly become a status round-robin: where's the deal at, what's blocking it, next. That was not a data point anyone tracked. It was just what happened when the math of one manager and eleven people met a Friday afternoon.

What Mindtickle's report does that nothing else has is put a real number on it, from inside actual usage data rather than a survey where managers self-report how much coaching they think they're doing. Three years of decline, 45% down, at companies who bought an enablement platform specifically to fix this. That is the part worth sitting with. This is not the neglected floor. This is the floor with the tooling already in place, and coaching still fell off it.

The interesting failure isn't that coaching declined. It's that a market-leading enablement vendor is reporting the decline as a success metric, because the thing that replaced the manager's attention was a copilot that reviews submissions, not one that runs sessions.

Is AI replacing sales manager coaching?

Not yet, and the report's own second stat shows why. Mindtickle says its AI reviewed 465,000 role-play submissions in 2025, work that would have taken one person sixteen months full-time (Mindtickle, 2026). That is genuinely useful automation. It is also a completely different job from coaching.

Reviewing a submission means scoring a recording against a rubric after the fact. Coaching means briefing a rep before the call on what good looks like, watching what actually happens, and debriefing the specific gap between the two, tied to the team's real methodology, not a generic scale. One of those is grading homework. The other is the thing that changes what a rep does on the next call. Mindtickle's own numbers show the grading scaled and the coaching kept shrinking, in the same report, from the same 400-plus companies.

That distinction is the whole ballgame right now, and almost nobody selling into this category says it out loud, because most of what's shipping is the first kind wearing the second kind's name.

What is the difference between an AI copilot and an AI coach?

A copilot assists a human who is still doing the coaching. An AI coach runs the coaching session itself, end to end: brief, observe, debrief, against your actual methodology, every rep, every time. Mindtickle's 465,000 submissions were reviewed by the first kind. Nobody ran the second kind at that volume, which is exactly why the 45% number didn't recover.

We built the second kind on purpose, because the copilot model has an obvious ceiling. A copilot that helps a manager review calls faster still needs the manager to show up and act on what it finds, and the whole point of this report is that the manager mostly doesn't. An AI manager doesn't wait for a human's calendar to free up. It briefs the rep, watches the roleplay or the live call, and debriefs it against the methodology, the same session quality at 8am on a Tuesday or midnight before a renewal. When SBR Consulting ran their capability accelerator on our system, every rep in the cohort got that full session, not a scorecard that landed in a folder nobody opened.

I've watched this play out from both sides: as the manager who ran out of Fridays, and now as the person building the thing that doesn't run out. The rubric-scoring version of AI coaching was never going to fix the 45% number, because it automated the part managers were already bad at skipping around, not the part that actually requires showing up.

What did Mindtickle's 2026 enablement report actually find?

Beyond the coaching decline, the report tracked 430% year-over-year growth in AI-authored content, with AI now generating 40% of role-play scenarios across the platform's customer base (Mindtickle, 2026). That's real signal about where the category is investing: content generation and scenario authoring, the supply side of practice. It says nothing about whether reps are getting coached on the deals they're actually working.

That gap between AI adding scenarios and AI reducing coaching is worth naming plainly. More scenarios to practise against is not the same thing as more coaching. It's more equipment in the gym, with the same empty spot where the coach used to stand. Mindtickle isn't wrong that AI is doing more. It's wrong, or at least incomplete, about which part of the job it replaced.

Does AI coaching actually change what a rep does differently?

That's the only test that matters, and it's a fair one to ask of us too. A copilot that speeds up review doesn't touch it, because speeding up a task nobody was doing consistently just means the same gap arrives faster. An AI manager that runs the full session, brief through debrief, against the real methodology, is built to touch it, because the rep leaves with a specific next action tied to a specific deal, not a score.

The honest version of this is that the 45% decline was never really about AI, up or down. It was a shortage of manager hours that existed long before any of this software shipped. AI didn't cause the drop. Mindtickle's report just gave it a number, three years late, from a company that sells the tool managers were supposed to be using the whole time.

Bring a real deal and see what actual coaching looks like

If you manage a team, you already know the 45% number in your gut, whether or not anyone measured it. The fix is not a copilot that helps you review faster. It's a coach that runs the sessions you stopped having time for, at the same standard, for every rep, every week.

Start free with Keenan at replicatelabs.ai, no card, no demo gauntlet. Bring the deal that's actually stalling and see what a coach that shows up every time looks like, not a rubric that scores what already happened.

Sources: Mindtickle 2026 State of Agentic Revenue Enablement Report