Sales coaching is one of the highest-leverage things a sales organization does, and one of the hardest to do well.
In our recent session with Mark Kosoglow (CRO at Docebo), Devang Agrawal (CTO at Airspeed), and Nicolas Galantini (Enso Connect), a few consistent challenges came up:
Managers are expected to coach across dozens of deals and hundreds of calls each week.
Most coaching runs on a small sample of interactions. And feedback, when it happens, is often subjective and hard to measure.
So coaching stays inconsistent, hard to scale, and rarely translates into lasting behavior change. That’s the real problem. Teams don’t lack frameworks or intent; they lack a system that makes coaching repeatable.
The data shows the same disconnect. 45% of reps now rate the coaching they receive as below average, up from 29% the year before. Yet 64% of sales leaders believe they are spending more time on coaching than 12 months ago.
Where Coaching Breaks Down in Practice
Most teams don’t lack coaching frameworks. They lack coverage, consistency, and follow-through. Mark Kosoglow described the reality most leaders face:
“You might have hundreds of calls in a week. Managers can only review a handful. So coaching ends up being based on a very small sample.”
That distorts the view of performance. Reps get coached on isolated moments, not patterns. Managers rely on what they happen to see, not what’s happening across the pipeline. And coaching is usually disconnected from execution.
- Feedback is given after the fact
- Training is delivered outside of live deals
- There’s little visibility into whether behavior actually changes
Which leads to a familiar outcome: activity increases, but performance doesn’t move in a predictable way.
The Shift: From Coaching as an Activity → Coaching as a System
The most useful part of the discussion was how teams are rethinking coaching. Not as something managers do occasionally, but as something the organization operates continuously.
Devang Agrawal put it plainly: “The issue isn’t that teams don’t know how to coach. It’s that coaching isn’t built into the way work happens.”
When coaching becomes a system, three things change.
1. Coaching Starts With a Defined Problem
Instead of broad feedback (“run better discovery”), teams isolate specific failure points:
- Deals stall after first meetings → weak problem definition or stakeholder alignment
- Pipeline looks full but doesn’t convert → poor qualification or next-step clarity
- Deals slip late → no mutual action plan or buying-process visibility
Nicolas Galantini shared how this changed their approach: “We stopped trying to improve everything at once. We focused on where deals were actually breaking - and coached that.” That creates focus and makes coaching measurable.
2. Leading Indicators Replace Gut Feel
Another consistent theme: most teams lean too hard on lagging metrics.
Win rate and revenue tell you what happened. They don’t tell you what to fix.
The teams making progress track behaviors inside deals and calls, such as:
- Whether a clear customer problem is defined
- Whether next steps are agreed and dated
- Whether multiple stakeholders are engaged
These indicators show whether reps are running the right motions, early enough to intervene. They also take the subjectivity out of coaching. Instead of “That call could have been better,” it becomes “There was no clear next step or buyer commitment. Let’s fix that.”
3. Coaching Moves Closer to Live Execution
Traditional coaching happens outside the flow of work:
- End-of-week call reviews
- Quarterly training sessions
The teams discussed are moving coaching into:
- Active deals
- Recent conversations
- In-progress accounts
As Mark put it: “The most effective coaching happens inside a real deal. That’s where context exists, and where change actually sticks.”
This takes a different level of visibility, and it’s where most teams hit a wall. Reviewing enough real interactions, consistently, doesn’t scale by hand.
Where AI Actually Changes the Model
AI didn’t come up as a replacement for managers. It came up as the infrastructure layer that makes this system possible.
It lets teams:
- Analyze every call, not just a sample
- Identify patterns across reps and deals
- Surface where coaching is needed without manual review
- Track leading indicators automatically
That shifts coaching from:
- Periodic → continuous
- Sample-based → comprehensive
- Manager-dependent → system-supported
Or as Devang summarized during the session: “AI doesn’t make you a better coach by itself. It makes coaching scalable.”
What This Means in Practice
The takeaway wasn’t to “do more coaching.”
It was to change how coaching is structured.
Teams seeing results are:
- Narrowing focus to 1-2 coaching areas at a time
- Defining clear leading indicators for those areas
- Embedding coaching into real workflows (calls, deals, accounts)
- Running a cadence of inspection (weekly, not quarterly)
And the part that matters most: they’re closing the loop between insight → action → verification. Without that loop, coaching stays theoretical.
The Real Advantage Isn’t AI: It’s Learning Speed
One idea came up again and again: the advantage isn’t access to AI. It’s how fast teams turn insight into behavior change.
The teams pulling ahead are:
- Instrumenting their sales process
- Identifying gaps faster
- Reinforcing behaviors continuously
Not perfectly, but consistently. Over time, that compounds.
