June 18, 2026 – 8 min read
by Performance Development Group, Coaching Excellence Practice Team
AI coaching tools are now standard in pharma sales: the simulators where reps practice their calls, and the systems that score and summarize them. As these tools take over the repeatable, correctable side of rep development, they leave first-line managers with a conversation that runs almost entirely on judgment, which is the part of coaching most commercial teams are least prepared for. This article maps what changes in the manager-rep one-on-one once AI is in the mix, and what leaders need to put in place so the technology strengthens coaching rather than erodes it.
The sales manager’s dashboard has never been more robust. Call summaries roll up on their own. Role-play scores, message-adherence flags, and next-best-action prompts arrive without anyone asking for them. AI-enabled tech has handed time back too: McKinsey found a generative AI prep tool gave more than 10% of seller time back, and Gartner projects that by 2027 nearly all seller research will start with AI.
For years the problem in pharma coaching was too little information, a quarter-old sales figure and whatever the manager happened to see on one field ride. Now there is almost too much. And across the engagements PDG runs, one thing has not moved with the data: the conversation between a manager and a rep is mostly the same one it was before.
The inputs to coaching have changed completely, but the conversation has not.
Put all that data in front of a leader and a predictable thing happens. They over-index on it and lose the part of coaching that was never in the data: sitting beside the person and seeing what they are actually experiencing.
For example, consider these two sales reps. The first individual is overachieving. Six months of metrics, every box checked. Look at her as a person, though, and she is slumped in her chair, volunteering for nothing, coasting on an easy territory and long-standing relationships. Strong numbers do not mean she has stopped needing coaching.
The second individual is missing his targets, so the dashboard says he is failing. He may be new to the role, or stuck in a territory where access is the real problem. The numbers paint one picture. You have to read the whole story to coach either of them well.
Fifty-eight percent of sales managers already struggle to get through what is on their plate in the time they have (Gartner). So when a tool arrives that looks like it has coaching covered, the temptation is obvious: the AI is handling it, I do not have to. That is the moment coaching stops altogether.
So if the coaching conversation hasn’t changed, what has? Practice. Before AI tools, a rep practiced maybe three times a year, at the January, mid-year, and year-end meetings. The rest of their practice happened live, in front of a physician, where a missed message is a missed opportunity. Now that same rep can practice 12 to 24 times a year on top of those meetings. That is a real gain in and of itself, regardless of whether you trust every score the tool produces.
This additional practice also changes what the rep brings into the coaching conversation: they arrive with the fundamentals already rehearsed, a clearer read on what landed and what didn’t, and specific moments they want help interpreting—not a messy first attempt they need the manager to clean up. The behavior the field has spent years trying to coach out of: reps walking in and reciting everything they know; the tools now catch. A rep cannot score well without building rapport and asking real questions first, fundamentals.
That is how AI tools free up the manager. Reps spend less than 30% of the week actually selling (Salesforce, 2024), and AI now absorbs some of the lowest-value parts of the manager’s job too, including the feedback documentation that used to pile up until Friday. AI handles the repetition. The manager handles what is left, which is harder.
An AI tool can tell you what happened on a call. It cannot tell you why. That conversation is the manager’s whole job now.
The conversation AI hands back is the one about why. A tool can tell a manager what happened on a call, the score and the step that got missed. It cannot tell them why it happened, and that is the whole job now. The managers who are good at it have kept a habit most of us drop after age four: when they get an answer, they ask why again.
The dashboard says a rep’s eye contact with the physician was poor. A manager treating the score like a report card hands it back: fix your eye contact. A manager who asks why gets more valuable information. Maybe she was uncomfortable. Maybe she was still reading the message off the screen because she had not internalized it yet. Each answer points to a different fix, and without the question, you can coach in exactly the wrong direction. Treat the dashboard data as a verdict and you coach to the result. Ask why and you coach the behavior underneath it.
This harder coaching conversation also carries the most risk. Handle it badly and you not only miss a chance to help, you do harm. In a survey of roughly 6,000 sales reps, Gartner found bad coaching sets performance back about twice as much as good coaching moves it forward. The same data that lets a sharp manager open a smarter conversation lets a careless one confidently coach the wrong thing. Bias makes it worse: a manager who already rates a rep as a star may wave off real feedback simply because of who it is about. Reading the whole person is human work. An AI tool can frame the question. It cannot answer it.
When a leader calls to say they just bought an AI sales coaching platform, the first question is about something else entirely: how the tool fits the rest of what the rep already experiences. Here is what goes wrong when no one asks. The rep sits in a January meeting about storytelling, gets back to the office to an AI prompt drilling questioning skills, then rides along with a manager focused on call objectives. Three priorities in one week, and the rep is left wondering which one actually matters.
When you align all of these activities, the same week works as one motion. Storytelling at the meeting, an AI persona who wants to hear a patient story, a manager on the ride reinforcing the same skill in live calls. The pieces fit instead of competing.
Frontline manager coaching is consistently the highest-leverage variable in sales performance (Gartner, 2024), so it is worth designing the system around rather than something you bolt a tool onto. PDG maps coaching as a five-stage flywheel: plan, engage, reflect, coach, reinforce. The rule is that it should keep flipping between machine and human. AI feeds the data and the practice. The human supplies the judgment, and the competency and performance calls that affect someone’s standing stay there, with a person who sees the whole experience and not a snippet of simulated practice. Get too heavy on the data and it leads you astray.
AI raised the floor. The basics get practiced before the manager ever sits down, and the data arrives clean. It also raised the ceiling, because what is left is the judgment behind the number, and that is the conversation that does the most damage when it is handled badly. The managers who win will use the freed time and the cleaner data to have that harder conversation, not to have the old one faster or to stop having it.
If a leader changes one thing on Monday, it is this. The field spends years telling reps not to throw data at a physician and expect to convince anyone. We should stop doing exactly that to our own people. Look at the number, then ask why.
The PDG Coaching Excellence Practice helps pharma and biotech commercial organizations turn first-line managers into effective coaches. Working exclusively in life sciences, the team designs the coaching systems and manager capability programs that close the gap between training and sustained field behavior.
COACHING The Pharma Mid-Year Review Should Be a Coaching Diagnostic. Here’s How. by Performance Development Group, Coaching Excellence Practice A pharma mid-year coaching diagnostic is a structured assessment that evaluates each sales rep against the demands of the next six months, not the past six. Unlike a standard performance review, which measures what already happened, […]
COACHING The Three Biases Distorting How Pharma Sales Leaders See Their Teams by Sean Frontz, in partnership with PDG Content Team Why do pharma sales leaders overrate their teams? Research shows 62% of performance rating variance comes from the rater’s own tendencies, not from the person being rated. The three most common biases in life sciences […]
COACHING The Misdiagnosis Costing Your Field Force by PDG Content Team in partnership with Sean Frontz Why does sales training fail to improve field force performance in pharma? Pharma commercial organizations consistently misdiagnose underperformance as a skill deficit, so they respond with more training. But the data tells a different story: roughly 90% of performance failures stem […]
