Feature
Earlier this year, Gallup crunched the data from over 100,000 employee engagement surveys and found “70% of variance in team engagement is determined solely by the manager.”
Great managers inspire, motivate and remove obstacles standing in people’s way. But too often, first-time managers transition from being a successful individual contributor into management with little to no instruction. And that lack hurts not only the manager’s motivation and performance, it hurts those people they’re managing and by extension, the broader organization.
Clearly first-time managers would benefit from stronger guidance and instruction when making this transition — and throughout their career. As with many areas of the workplace, GenAI-based solutions from companies like Wisq and Valence are entering the market as a potential solution to the problem. But how far can GenAI go in teaching what we might consider fundamentally human skills?
Is a GenAI Teacher Better Than No Teacher at All?
Something is better than nothing, Boston Consulting Group’s Debbie Lovich told Reworked. She categorized the training offered first-time managers as “woefully insignificant.”
The relationship between managers and team members is where an employee’s experience is made or broken, which makes that initial transition from individual contributor to people manager all the more important.
“There are so many little things in being a manager that can make a big difference,” Lovich said.
The lack of training is an argument in favor of using GenAI as a support system, agrees neuroscientist, executive coach and workplace culture advisor Jeanine Stewart. But she adds, “I do think we need to be really thoughtful about the limitations.”
“One thing AI and AI-supported training is helping us to do, is giving managers who have often been thrown into this career path with no training … access to some really good content. And you can modify it with some tools to meet you where you are, which is remarkable,” Stewart said. AI in this case, serves effectively as a launching pad for these new managers.
Julie Starr, coach, author and founder of Starr Coaching Ltd., also sees GenAI serving a purpose in this scenario — with limitations. “There are questions that would form useful tools to put a manager through to have them decide, ‘What's the criteria of success for my role?’ ‘What must I do to be judged as high performing? It doesn't take a human being … there are questions that you could put any junior manager through. And they would be useful,” Starr said.
But she makes a distinction between training a manager versus coaching, which comes down to the difference between directive feedback and productive inquiry. The former is helpful to set a foundation and establish common frameworks, but the latter is needed once a certain level of knowledge has been obtained.
Given the existing frameworks and knowledge available in the coaching world, Starr believes AI “could usefully augment, or even in some cases, replace the basic offering of coaching." She views it being competent at a rudimentary level, the equivalent of someone who has "read a few coaching books, been on a couple of courses, I've still got stabilizers on my bike."
GenAI’s abilities to ingest, consolidate and quantify large quantities of data is where Lovich sees it holding promise. She speculates about a potential application in a globally distributed organization of 100,000, where GenAI could codify the practices and behaviors of the leaders who are identified as the best mentors, or the strongest people connectors, or the best coaches. “If you could find all those practices, and bring them together, as to a set of best practices at distributed team management,” GenAI could then be used in that case to reward the positive behavior and give prompts or nudges in real time where correction is required, she said.
Related Article: AI-Powered Career Development Brings Great Potential. It Also Brings Ethical Risks
Learning Is Experiential and Observational
Discussions of GenAI’s potential typically revolve around a vague promise of “freeing people up to do more valuable work.” In the case of using GenAI to teach first-time managers, the value proposition becomes more about the potential for scale and access.
But there is a cap to how far a manager can — or should — go before human intervention. As all three say, learning is experiential and observational. While interactions with a personalized GenAI guide can provide useful in-the-moment insights or nudges, it cannot replicate the power of observing someone else's model desired behavior.
Stewart acknowledges that this challenge extends beyond the context of AI and applies to the entire time period since working from home became the norm. “One of the losses of this era has to do with the lost ability to watch people model the right way. And this is true in training as well,” she said.
What also gets lost in the human-GenAI training scenario is what a master coach accomplishes through observation of their client, said Starr. “As a coach with thousands of hours of experience, you learn to work with your logical mind, your intellect, and also your less logical mind. Maybe insight, intuition, wisdom, but also you work with your whole body. And I do mean your physical body,” she said.
So while she sees generative AI as a good supplement and potentially even a replacement for some of the rote learning which supports an entry-level manager, she believes this will be a limit to how far GenAI can go unless it becomes sentient. Starr questions if it will ever be capable of picking up on emotions expressed through non-verbal cues.
Stewart shares a similar point. She’s watched first-hand as leaders have come to recognize the extra effort needed to show up effectively when in virtual settings. Implicit in this statement is the additional barrier created when that interaction is between human and machine. She believes leaders face the challenge of needing to create what she calls “neural synchrony,” to help build a better sense of connection. “All of the nonverbal and subtle hints that I am with you, I’m listening to you, I’m supporting you.”
Of the three, Lovich sees the greatest potential for GenAI for manager training, but she also acknowledges its limitations. “Learning is experiential and observational. Having a coach who could see what you're doing can get you to practice different things and get you to reflect on it.” While she admits AI will eventually be able to do that, she still wishes “the initial training to be more experiential, more hands-on, more human,” with GenAI acting as a nudge or as a scaffolding to support.
Related Article: Middle Management Is Stuck. Time for a Reset
AI's Limitations Once You’ve Gotten Rid of the Training Wheels
Large language models are trained on existing information — either what it’s been fed, in the case of the custom-built coaching tools, or information it gathers from an organization’s system when using less-focused, commercially available models for coaching. Given the upheaval in most workplaces over the last four years, Stewart notes in the latter case in particular, the language models would be drawing from older insights, not necessarily attuned to the current reality. Because as she said, “There are no managers out there who have a depth of experience managing and leading in this particular era.”
Conference
Sep
16
HR Tech Conference Las Vegas 2025
Register
Conference
Oct
27
Gartner HR Symposium/Xpo Orlando 2025
Register
Starr notes the changing needs of managers as they gain more experience.
“As soon as a manager starts to get some experience and starts to understand the complexities and the different dynamics of people, process, technology, strategy, context, culture — as soon as a manager starts to understand all these different dynamic forces, it's gone beyond a little bit beyond logic, hasn't it?” she said.
At that point, a human coach should step in who can parse the nuances feeding performance and results. Starr notes the external factors that shape all worker’s performance and how it could be challenging for AI to parse the best responses given the multiple factors as such responses don’t necessarily exist in so-called “best practices.”
When asked if AI’s notorious hallucinations are a problem, Stewart was pragmatic: “Look, we’ve always had a mixture of good and bad advice flowing, right?” So the challenge becomes how leaders can guide employees in evaluating the output of the tool and how to understand the root cause in cases where it is false or inaccurate. That would require leadership having oversight and the ability to tweak these models once the origins of potentially bad leadership practices are identified.
Quality control is a concern for Lovich, too. “You've got to run a lot of simulations of AI … What advice is it giving? You wouldn't put something out to customers that you didn't quality control. In the same way, you shouldn't do that with your employees. It needs the same risk management and A/B testing.”
Related Article: Use Generative AI to Write Performance Reviews? Not So Fast
We Always Return to: Keep Humans in the Mix
One of the questions we’re asking when looking at the application of generative AI in the workplace — and in our lives — is if it is increasing our agency and autonomy or diminishing it. Or as Stewart asks, “How do we make sure that what they’re executing on is adding value and not diminishing the motivation and inspiration of the people you need to create business impact?”
In the case of GenAI helping train first-time managers, particularly given the dearth of support, it feels as if it is adding value, provided we apply good governance, and recognize its capabilities and limitations.
And of course, as in all cases of generative AI's use, the caveat applies: keep humans in the mix.
