Most B2B sales teams sit at Level 1 of AI-assisted coaching (efficient call review) and are working toward Level 2 (targeted skill coaching). Very few have reached Level 3 (a fully owned intelligence layer), and Level 4 (real-time, in-call AI coaching) is still aspirational industry-wide.
Every sales leader knows the coaching math. An hour a week of skills dosing, roughly 3% of working time, produces 15 to 20% performance gains, a finding replicated across multiple studies and popularized by an HBR piece often called “the dirty secret of sales coaching.” And yet 77% of teams say they don’t do enough of it.
The gap isn’t a lack of will. It’s bandwidth. A frontline manager running a quota, a team, and constant reporting simply doesn’t have the hours to watch every call and turn it into personalized development.
That’s the gap AI is starting to close, but not all at once, and not the same way for every team. In a recent Winalytics-hosted peer roundtable, sales and revenue leaders compared notes on exactly how far they’ve each gotten. Their experiences map cleanly onto a four-level maturity model.
The Four Levels, at a Glance
- Efficient call review: using AI to surface the calls and moments that matter, without adding to a manager’s week.
- Targeted skills coaching: rubrics that tie specific, coachable skills to specific moments, so managers know exactly where to spend their time.
- An owned intelligence layer: conversation data connected across CRM, training, and go-to-market motion, so insight turns into action without manual work.
- Real-time and simulated coaching (aspirational): AI role-play and live, in-call coaching. Nobody on the call had fully arrived here.
Level 1: Efficient Call Review
Mark Guthrie leads a matrixed sales team at Verisk and has been living inside Gong for about a year. His take on what changed for him as a leader: being able to “witness and see the customer calls and interactions with our team members is extremely valuable… it’s a game-changing thing for me.”
His team built a custom AI-scored brief specifically for objection handling, checking five things on every call: did the rep acknowledge the objection, understand it, handle it clearly, ask clarifying questions, and confirm it was resolved. They’re also tracking something less obvious: whether reps open calls with a clear, legitimate purpose statement, and whether that habit trends up over time across the team.
His advice to teams just starting out: don’t try to instrument everything at once. “Pick three things… guide your AEs based on the briefs and the scoring, that they can assess themselves… spoon-feed them as you climb the ladder.”
Level 2: Targeted Skills Coaching
Brent Keltner has seen this pattern repeat across the teams he works with directly. In one engagement, his team built eight distinct rubrics inside a call-recording platform, each tied to a specific, coachable skill: getting to a real discovery impact statement, running a qualification framework like MEDDIC, structuring a micro-presentation. Reps tag their own calls to a rubric, and managers get a skill-by-rep view that shows exactly where to spend limited coaching time.
“Most of us are really at level one, right? More efficiency in call reviews,” Brent said, describing where most of the group’s own teams currently sit. The jump to Level 2 is about turning that efficiency into precision: instead of a manager spending an hour and a half or two hours a week skimming calls across 18 people, they’re reviewing tagged, scored highlights and giving feedback that’s “much more targeted.”
Level 3: Owning the Intelligence Layer
Aaron Verasammy, Chief Sales Officer at Mevo, described a setup a level ahead of the rest of the group, and was candid that it took real, sustained investment to get there. The core decision his team made: rather than depend entirely on one vendor’s platform for both call intelligence and coaching insight, they built their own data warehouse underneath it, so the underlying context and intelligence stays in-house. “We want to own the context, and we want to own the intelligence,” as he put it.
The payoff shows up as a daily “command center”: every manager gets a morning email that scores each call, flags key objections, checks whether the CRM was actually updated, and tells them exactly where to focus that day. Reps get their own version ahead of weekly one-on-ones, so nobody walks in without having already reviewed the data.
In Aaron’s case, this is not just a data exercise but instead one that is driving meaningful sales impact, with improvement across the board, from increasing outbound volume to higher close rates and shorter cycle times.
Maybe the most telling insight had nothing to do with the technology itself but instead the gap between what a prospect says and what a sales rep hears. “This was the aha moment for me… [a rep says] ‘oh no, we got no objections, man.’ But the AI picked up four objections. Now my managers can step in and ensure that our reps are ready to address those objections on the next call.” Owning the intelligence layer didn’t just save managers time. It surfaced blind spots reps didn’t know they had.
Level 4: Real-time and simulated coaching (Still Aspirational)
Nobody on the call claimed to have fully arrived at Level 4, and Brent was upfront that it’s aspirational for the whole group right now. Two ideas came up as the likely next frontier: AI-based role-play for deliberate skill practice, and in-call, real-time coaching, an AI advisor present alongside the rep during the conversation itself rather than reviewing it after the fact.
Where Most Teams Actually Are (and Why That’s Fine)
If there’s one honest takeaway from the conversation, it’s that even sophisticated B2B sales orgs are mostly still building Level 1 habits. Tess Jenkins, VP of Sales at Campus ESP, was refreshingly candid about it: her team is still building scorecards and building the discipline of consistent call review, describing her current process as “very manual” by comparison to where others in the room had gotten. That’s not a lagging indicator; it’s the normal starting point. The value of a maturity model isn’t in shaming teams for being at Level 1; it’s in giving them a clear next step instead of a vague mandate to “use AI more.”
The Takeaway
You don’t need to leap to Level 3 to get value. Most of the performance gain is unlocked simply by moving deliberately from Level 1 to Level 2: from reviewing calls efficiently to reviewing them against skills that actually matter for your team’s growth and revenue targets. Know your level, then invest in the next one.
