4 Ways Sales Managers Turn AI Sales Coaching Into Lasting Behavior Change

Once reps have AI sales coaching, the hard part of coaching shifts from finding feedback to choosing which piece of it to act on. Depending on the tools, AI can score recorded calls, run role plays whenever a rep wants them, and deliver feedback minutes after a call or practice session ends. That can leave a rep with a long list of accurate notes and no clear sense of which one to work on first.

A manager who uses the tool to pick that one change, and then coaches it, turns the list into one behavior the rep changes. Here are 4 ways sales managers use AI sales coaching to change rep behavior, plus how to split the coaching plan between the manager and the tool.

Key Takeaways

• AI sales coaching gives managers more to work with: Call-analysis tools score recorded calls, and AI practice tools run role plays on demand, both with instant feedback.
• Choose the one behavior worth coaching: Start from where the rep's deals stall, and let the other flags wait.
• Find out whether the rep can't or won't: A rep who can't needs practice, and a rep who won't needs a different conversation first.
• Give the rep a reason to change: Tie the change to something the rep wants, like a promotion or a bigger book of business.
• Hold the rep to a plan they wrote: The rep drafts the plan, and the manager stays with it until the behavior holds on real calls.
• Split the coaching plan on purpose: The manager directs the coaching, and AI handles the repetition.

What AI Sales Coaching Does Well, and What It Takes to Change Rep Behavior

AI sales coaching handles the high-volume parts of coaching that no manager could cover alone. The tools generally come in two types, and many teams use one or both:

• Call-analysis tools: These review recorded customer calls and flag patterns, such as how often a rep asks about the decision process or confirms a next step. A manager can listen to a handful of calls, and these tools can cover all of them.
• AI practice tools: These run role plays with an AI sales coach playing the customer, so a rep can practice the night before a big meeting, as many times as they want.

Both give reps specific feedback while the call or role play is still fresh.

Reps are also using general AI tools on their own, asking for feedback on a message or a talk track before a manager ever sees it. Some of this coaching used to wait for a 1:1, and much of it never happened at all.

AI feedback is only part of what changes a behavior, though. A rep changes a behavior when three things are true:

• They're focused on one thing: A single behavior, not a list.
• They want to fix it: The change matters to them personally.
• Someone holds them to it: A person checks in and keeps them on it.

Both types of tool cover the feedback well, and many can rank what they flag. What they don’t know is which behavior is the biggest factor costing this rep deals in these accounts, why the rep hasn't changed it already, or what would make this rep want to.

Take a call-analysis score after a rep's first meeting. Let’s say it flags a high talk ratio, a skipped budget question, two filler phrases, and a vague next step. The rep tries to fix all four on the next few calls, and a week later none of them has stuck. Each note was accurate. The problem was that a comprehensive list isn’t that helpful.

Gartner sees a similar limit with AI tools in sales more broadly. It predicts that by 2028, fewer than 40% of sellers will report that AI agents improved their productivity, and warns that adding more prompts and tools to sellers' workflows risks overwhelming them.

Gartner also makes a related point about where the sales manager role needs to go. In its guidance for sales leaders in the age of AI, it calls for redesigning the manager role so managers act as amplifiers of seller effectiveness, and notes that CSOs often underestimate how much managers need the same kind of support and development their sellers get.

For enablement, one practical way to give that support is to be clear about the four coaching jobs managers lead, and how AI fits into each one.

#1: Choose the One Behavior Worth Coaching

The behavior worth coaching is the one behind the customer outcome the rep keeps missing. AI feedback shows a lot about how a rep handles a conversation, and the manager's job in coaching sales reps is finding which of those behaviors is costing deals. That search starts on the customer's side of the conversation.

Each point where a deal stalls points to a different part of the rep's conversation:

• The customer won't agree to meet: Look at how the rep opens, and whether the outreach leads with the customer's problem.
• The customer won't share what's driving the decision: Look at discovery, and whether the rep follows up when a first answer is vague.
• The customer won't embrace the recommendation: Look at whether the rep ties it back to what the customer said they need.
• The customer won't commit to a next step: Look at whether the rep asks for a specific next step with a date on it.

Tying each behavior to the customer outcome it produces turns a long scorecard into a short list. From there, three checks narrow it to one:

1. Find the pattern: Look across several of the rep's recent deals for the stage where they stall most often, since one lost deal doesn't make a pattern.
2. Filter the AI feedback: Keep only the call or practice feedback connected to that stage, and let the rest wait.
3. Listen to a call at that stage: Hear how the rep handles that moment, with what the manager knows about the account in mind.

Say a rep's proposals keep going quiet. The team's call-analysis tool flags long monologues and a fast pace across several calls. Listening to two of the rep's discovery calls shows something the scores don't: when a customer gives a vague answer about what's driving the purchase, the rep accepts it and moves on. The proposal gets built on surface answers, so the customer has no reason to act on it. The stall showed up at the proposal, and the behavior behind it sat earlier, in discovery.

That third check is where the manager's judgment matters most. A call-analysis tool scores what happened on the call, and a practice tool scores the role play. The manager knows the account's history, who else is weighing in on the decision, and whether this customer was open to the rep before the meeting started, and that context can change which behavior is worth the coaching time.

#2: Find Out Whether the Rep Can't or Won't

AI feedback can show that a behavior is missing, but the manager is the one who needs to find out why. A rep who can't do something needs practice. A rep who won't do it needs to have a different conversation first, because more practice won't help.

Three signals sort one from the other:

• Whether the rep can do it in practice: If they handle the moment well in an AI role play or a role play with the manager, the capability is there.
• Whether the rep does it on live calls: Strong in practice and missing on live calls points to "won't," or to nerves on high-stakes calls. Weak in both points to "can't."
• Whether the rep does the practice they agreed to: Desire shows up in what a rep does, so completed practice says something different from practice that keeps getting skipped. When practice runs through an AI practice tool, completion is easy to check.

With that said, skipped practice can also mean a rep buried in end-of-quarter deals is simply out of hours, so it's worth asking questions before drawing a hasty conclusion.

Say a rep handles vague answers well in AI role plays but skips the follow-up question on live discovery calls. The capability is there, so more role plays won't fix it. The manager's next move is a conversation about what's different on a live call.

When the answer is "won't," the cause tends to fall into one of four places, and each calls for a different response:

• The relationship: Something is strained between the rep and the manager. The manager addresses that first, including owning their part of it, because coaching rarely lands until it's settled.
• Ego: The rep struggles to admit they need to improve, or to accept help. The manager makes clear they're on the rep's side and keeps the expectation for results clear.
• A competing belief: The rep thinks their current approach already works. Asking them to test the new approach on a few calls and compare results tends to move them further than arguing.
• No clear payoff: The rep can't see why the effort is worth it. Asking what they want from the next year, and connecting the change to it, gives them a reason.

Separating can't from won't first keeps a manager from drilling a capability the rep already has, which matters most with a rep who has quietly stopped trying.

#3: Give the Rep a Reason to Change

A rep works on a behavior when they can see what's in it for them, and the manager is the one who makes that connection. Without that connection, a scorecard can put a rep on the defensive, the same way a product pitch puts off a buyer.

That connection starts before the 1:1. The manager needs to know what each rep wants, personally and professionally, which comes from asking directly and paying attention over time. Personal goals, like buying a house or moving into a bigger role, often matter more to a rep than the team's number.

In the coaching conversation itself, a few moves bring the AI feedback and the rep's goal together:

1. Let the rep go first: Ask how they think the call or role play went and what they'd change. Reps are often harder on themselves than a scorecard, and their answer shows what they already see.
2. Start with what's working: Name what they did well, so they know what to keep doing.
3. Ask your way to the gap: Start broad and narrow down. For a rep who accepts vague answers in discovery, that might sound like "What did you learn about what's driving this purchase?" then "What else would have helped you build the proposal?" then "What did they say mattered most to them?" Reps rarely argue with a problem they named themselves.
4. Connect it to their goal: For example: "Because you want to move into strategic accounts next year, getting to the real driver in discovery is what will set your proposals apart."
5. Check for hesitation: Before wrapping up, ask whether anything is holding them back from trying it on their next call.

The expectation for results stays firm, and how the rep gets there is their call. Reps take this kind of feedback best from a manager they trust is in their corner, and that relationship is what makes any feedback worth acting on, whether it comes from the manager or the tool.

#4: Hold the Rep to a Development Plan They Wrote

Accountability holds best when the plan belongs to the rep. Drafted by the rep at the end of the coaching 1:1 and agreed to by the manager, it becomes the rep's own commitment, rather than one more item on the manager's to-do list for end-of-quarter forecasting to bury.

It can fit on half a page:

• The behavior: The one thing the rep will change, such as asking a follow-up question whenever a customer gives a vague answer about what's driving the decision.
• The practice: Two or three specific activities, such as AI role plays on that moment, listening to a call where a teammate handles it well, or writing down three discovery questions before each meeting.
• The measure: How the two of them will know it's working, such as the behavior showing up on the rep's next several live discovery calls.
• The check-in: When they'll look at progress together, with a date on it.

Holding the rep to the plan is mostly about what happens at each check-in:

• Open with the plan: Review the practice the rep committed to before anything else in the 1:1.
• Pause when the practice didn't happen: If the rep skipped it, the coaching waits while the manager finds out why, since that's a can't-or-won't question again.
• Expect early attempts to fail: The first live attempts will be clumsy, so the manager keeps the stakes low and the tone encouraging.
• Move on only when it's consistent: The behavior is ready when it shows up call after call, and one strong call is only a start.

The rep does the work, and the manager keeps them on it. The common mistake is moving on to the next thing the AI flagged before the first behavior holds on real calls.

How to Split a Sales Coaching Plan Between AI and the Manager

An AI-enabled sales coaching plan works best when the manager directs it and AI handles the repetition. Enablement can build that split into the rollout of an AI coaching tool, so managers know how the tool fits into their coaching.

Here's how the work divides in practice, and how to tell each piece is working:

• Call review: Call-analysis tools flag patterns across recorded calls, and the manager picks the one behavior to coach and listens to real calls. On teams without call analysis, the manager's own listening covers this. It's working when the coaching focus matches where the rep's deals stall.
• Practice: AI practice tools provide role plays and repetitions on demand, and the manager assigns practice tied to the chosen behavior. It's working when the rep completes the practice they agreed to.
• Feedback: Both types of tool deliver specific notes right after a call or role play, and the manager connects the change to what the rep wants. It's working when the rep can explain why the change matters to them.
• Development plan: AI feedback from calls and practice supplies the evidence the plan draws on, and the manager agrees to a plan the rep drafts. It's working when each rep is focused on one behavior at a time.
• Follow-through: AI practice tools track completion, and the manager follows up in the 1:1 and stays on the same behavior. It's working when the behavior shows up on live customer calls.

In each case, AI makes the work visible and repeatable, and the manager decides what it means for this rep. That works best when the split lives inside the workflows managers already run, instead of becoming one more dashboard to check.

How to Tell Whether Managers Are Getting the Most From AI Sales Coaching

Managers are getting the most from AI sales coaching when a rep can name the one thing they're working on, why it matters to them, and where it showed up on a real call. If the rep can only point to their latest score, the coaching is stopping at the scorecard.

A few checks make this visible for an enablement leader:

• The 1:1 adds something new: The conversation covers what the score can't, like account context or the rep's own reasons, instead of reading the scorecard back.
• The rep knows their focus: They can name the behavior they're working on without opening a dashboard.
• Managers hear real calls: Each manager has listened to at least a few of the rep's calls, beyond reading any AI summary.
• Coaching has its own time: Managers hold protected coaching time for development, separate from deal reviews.

When these hold, the AI is doing what it does best, and managers are using everything it surfaces to coach one behavior at a time.

Build Your Sales Managers’ Coaching Capabilities

AI sales coaching gives reps more feedback and practice than any team has had before, and sales performance coaching from their managers decides what that feedback adds up to. One quick way to gauge where your team stands is to sit in on a few 1:1s and ask: did this conversation add anything the scorecard didn't?

If it didn't, that's a gap managers can close. Picking the right behavior, telling can't from won't, and steering a conversation toward a plan the rep owns all improve with training and practice, the same way selling does. It's the kind of support Gartner says CSOs tend to underestimate.

Catalyst™ equips frontline managers to diagnose capability and desire gaps, work through the barriers that hold people back, and build development plans with each rep. Schedule a consultation to see how it fits alongside the AI tools your team already uses.