Traditional coaching has always had a significant bottleneck: One skilled leader could have a team of 15, 30, sometimes 40 people. No matter how good the coach, there’s only so much time in the day. Most companies have accepted that constraint as a fixed cost of doing business.
That’s changing with the arrival of AI coaching. But should businesses focus on one over the other? Not necessarily.
Research suggests that both approaches work. Each has its own strengths, but in many cases, they complement each other rather than compete. Where AI coaching makes daily practice accessible and scalable, human coaches bring the judgment that complex, high-stakes moments demand.
In a recent BetterUp study, 51% of employees said they’d be interested in a hybrid of both.
The real question for business leaders is which one to deploy, and when, so you amplify the outcomes.
Human Coaching At Work
Human coaches bring what no algorithm can replicate. They can sense the politics of a reorg, find the right tone for a hard conversation, or help someone think through a career decision they haven’t fully put into words yet. That kind of judgment takes human expertise, which comes from lived experience and an instinctive read on the people involved.
The trick is, human-to-human coaching doesn’t always scale easily, and it isn’t cheap. Take TEKsystems, which hires around 350 new sales development reps a year, most of them stepping into complex, global IT sales with limited experience. Getting those reps ready to talk to customers used to depend almost entirely on sales coaching time with their direct leaders for roleplays and pitch reviews. But the leaders were each managing 15 to 40 direct reports across more than 100 locations. The cost of their expert time was high and there simply weren’t enough hours in the day to give every new rep the coaching they needed to gain confidence quickly.
Expert time is only getting more valuable. Every coaching moment needs to count.
AI Coaching At Work
AI coaching today is a mix of conversational guidance and structured practice. A sales rep runs a pitch against an AI coach that responds like a real prospect, pushes back with objections, and gives immediate feedback on tone and clarity. A new manager talks through a tough feedback conversation with an AI coach, then gets pointers on what to say differently next time.
That’s exactly what makes it useful for daily practice. It’s available at 11 p.m. the night before a big presentation or ten minutes before that really important client call. It doesn’t get tired of running the same roleplay ten times in a row, and it gives feedback right away instead of at the next scheduled check-in. There’s more freedom to fail without judgment, no matter how poor the attempt was.
Employees appreciate the accessibility and added psychological safety. In a recent Conference Board survey, 91% of employees who’d tried AI-based coaching said they’d use it again, and 96% said the coaching felt personalized to them.
The next frontier is proactive coaching, or AI that reaches out before you need it. Degreed’s AI Labs team is testing this with Invisible, an experimental AI coach that messages someone ahead of a high-stakes moment, like a big meeting on their calendar, to prompt practice and discussion.
What AI coaching still can’t do is replace judgment built on human experience. It can help someone rehearse a hard conversation, but it can’t read the room during the real one.
The Hybrid Coaching Formula
The principle is straightforward: Use AI for volume and repetition, and reserve human coaching for the moments that require real leadership and expertise. In practice, that’s taken different forms at different organizations:
TEKsystems used Degreed Maestro to close the coaching-capacity gap that lengthened sales onboarding time. Reps got a place to rehearse before a leader was ever in the room, so leaders stayed focused on developing people instead of covering the basics. The program included four AI-coached conversations, each running roughly 15 minutes.
« Overall, we found astonishing feedback. Their confidence grew. They were able to roleplay in a safe environment, and that meant not having a leader or a mentor over your shoulder listening in. But they were able to do the Maestro experiences multiple times over until they felt that they were good enough to then go ahead and approach that call or test out their ability, » said Stefanie Kuehn, Senior Program Manager, Organizational Development at TEKsystems.
By the time a new rep sat down with a leader for a live coaching session or deal review, they’d already logged their reps. Leaders could skip the fundamentals and spend their limited time on refinement, which is where their expertise mattered most.
ZS, a global consulting firm, built the same principle into its own rollout. Jennifer Sutherland, ZS’s Global Head of Learning Enablement, calls the company’s internal AI coach « human enabled, » with AI running behind it, on purpose.
« You don’t want a system or a tool or technology to give you feedback. You want it to be human enabled, » Sutherland said at Degreed’s LENS conference.
In both cases, AI coaching earned trust through practice, feedback, and repetition. Human coaching was the last mile, where the real refinement happened. TEKsystems and ZS both kept that line clear on purpose, which is likely why the confidence and buy-in in both stories came so fast.
Match the Coach to the Moment
The question was never whether one form of coaching should replace the other. The skill leaders need now is knowing when to hand off the work to an AI coach like Maestro and when to bring in a human.
AI handles the early repetitions, roleplays, and on-demand guidance well. People are irreplaceable for in-depth expertise, fine tuning, and relationship building. Get that right, and coaching becomes a system that gets people ready faster, and keeps them ready as the work keeps changing.
Restons en contact
Je souhaite m’inscrire à la newsletter mensuelle pour me tenir informé(e) des prochains événements, bénéficier d’éclairages exclusifs et recevoir l’actualité des solutions Degreed.