Sponsored Article
White Paper
The Future of Leadership Development: AI-Driven Manager Enablement
Employee reluctance around AI usage is very real, although we saw similar headlines in the 1960s when workplace automation was on the rise. Eighty years ago, people were worried technology would replace them, just as people have the same fears around AI today. Back then, automation shifted productivity expectations, while today, AI will augment workers to shift what is possible to do with data. With a thoughtful groundwork and clearly identified metrics and goals, AI coaching has the potential to elicit real behavior change in the workforce.
AI coaching can be part of a larger effort to adopt more AI tools as organizations look for ways to enact behavioral change and boost employee engagement and workplace culture. According to Reworked research, AI investments are a significant priority for 73% of organizations. Employee concerns can be addressed with a thoughtful approach to AI coaching: understanding how it should be used, how to deploy it and what needs to happen for it to be truly effective.
To get a deeper understanding on AI coaching and what its best use cases are, Reworked interviewed Kirsten Moorefield, co-founder and chief strategy officer for Cloverleaf. We discussed Cloverleaf’s philosophy on what AI does best regarding coaching and how organizations can use AI to elicit real behavior change, rather than just rack up login metrics on an app.
Why Is Workplace Friction a Good Thing and How Is AI Disrupting Friction?
There’s been a decades-long societal trend of turning to technology, rather than coworkers, for connection. AI in all its forms has accelerated this trend, often with an air of sycophancy as the answer engine or chatbot agrees with everything you say. This technological dependence strips away the "friction" required for creating genuine human relationships and driving workplace innovation.
This, Moorefield argues, will set organizations back if they don’t address improving their culture of communication. Bringing conflicting perspectives together is exactly what generates new and better solutions. “What you need for true innovation at work is friction,” Moorfield said. “You need somebody who's in sales or product who has competing goals, and you need to come together and form a wholly new solution that neither of you could create on your own. You can't do that if you don't have the ability to be disagreed with."
Improving the gaps in soft skills is just one area where AI coaching can improve workplace culture. But not just any AI coach. It takes an AI coach specifically designed to identify patterns in behavioral data.
You don’t need an AI to ask how you’re feeling. For one thing, this is what human coaches are for. Further, using an AI model that emulates a human coach is both inefficient and off-putting to employees.
Why Does Human Mimicry Limit AI’s Potential? What Should AI Be Doing Instead?
There’s a prevailing trend among AI apps and tools to be given a name and an avatar, in an endeavor to be more human-like. But this can backfire. Recent research suggests that while AI can match human coaches around knowledge transfer, employees themselves aren’t too keen on AI acting like a therapist, as AI lacks genuine emotion and opinions.
Moorefield argues that trying to make AI act this way also limits the technology from its full potential. “We shouldn’t build our AI coaches to act like humans,” Moorefield said, “because it’s not leaning into the core strength of what AI can do better than humans.”
Because AI can process complex datasets faster than humans can, Moorefield believes this is the task we should set our AI coaches to, while leaving empathy to human coaches. "AI is best at sifting through all the data you don't have time to go through, giving you the one insight you need for the problem you're facing right now so that you can move forward with more confidence," Moorefield said.
Cloverleaf has made the clear design choice to not anthropomorphize their tool. Its AI tool is unnamed and ungendered, with no avatar or face. “Our AI will only use first-person pronouns for process-oriented things, such as ‘I’m pulling the data.’ What it won’t do is use ‘I’ or ‘we’ to display human empathy. Because AI is a math equation. We’ve seen studies time and again that humans reject emotional feelings an AI tries to impart, so our belief is it shouldn’t ever say something like ‘I’m proud of you,” Moorefield said.
How Can AI Deliver Value in the 'Flow of Work'?
Fragmented or disconnected tools are a great source of frustration for employees, and an impediment to adoption. Only 19% of organizations say their digital workplace is fully productive, according to Reworked’s most recent State of the Digital Workplace report. For AI coaching to be effective, it shouldn't require employees to log into a separate platform or spend long periods interacting with a bot.
Instead, it should meet employees where they are and provide answers while being as discreet as possible. Employees want AI to answer their questions so they can get along with their workday. “It has to come to people,”” Moorefield said. “And it can't just come to them with, 'Hey, what are you thinking today?' It has to come with, 'I know what's happening with you, and here's something that you probably need.'" This means HR leaders should be tracking more than login metrics to see the effectiveness of their tools.
These days, HR tech is very siloed, with the employee engagement survey living in one tool that might not connect to the assessment tool or the performance review tool. To improve adoption, AI coaching must sit on top of integrated, cross-system data, not isolated tools. This way, AI can pull from any data inside the overall ecosystem, instead of being limited to data from an isolated tool.
Engaged, But Not Convinced: What's Eroding Beneath Your Strongest Engagement Numbers
Watch Now
Retention Is a Communication Problem: New Data from 6,000+ Employees
Watch Now
Because starting from good data is where AI coaching succeeds.
What’s Needed to Start With AI Coaching?
AI coaching can’t be successful if it starts from "zero context." AI tools must be built on deep behavioral and relational data. Organizations looking to use an AI coaching tool (or build their own inside their ecosystem) must provide the tool with relevant data points. Simply waiting for people to use the tool and have the tool learn as it goes doesn’t work for this type of coaching model.
Moorefield offered an example from within Cloverleaf, where two employees were able to use the coaching tool to solve a culture issue. “One of our product leaders needed to tell our CEO that his team was stretched too thin. An aggressive deadline was putting quality and retention at risk. But when these two talk, the product leader will frequently soften his message, because he values the working relationship more than he wants to be blunt about what he needs. And to our CEO, a soft message makes it seem like the information isn’t entirely important.
“At the heart of the issue was that neither understood the relationship dynamic. All the product leader knew was that he'd failed to effectively communicate with the CEO in the past. But he didn't know why. So, our product leader asked Cloverleaf how to approach our CEO in a way that would make his issue be heard. And because the tool knows the personality and communications profiles of these two people, it was able to suggest a solution that would make each other hear and understand the other. This was game changing for their relationship and how they communicated with one another.
“And it happened because Cloverleaf had all the data it needed. Cloverleaf uncovered the product manager’s tendency to soften messages in protection of the relationship, and how our CEO receives that softening as a signal the information isn't critical. Neither party was doing anything "wrong," just different. They needed help figuring out what to adjust to effectively understand each other and the situation, and then find a solution. This kind of epiphany, followed by specific advice that actually lands, is the power of the AI Coach trained on the right, individualized data.”
Frequently Asked Questions
Workplace friction (defined here as navigating conflicting goals or disagreements between individuals or teams) causes people to think about solutions they wouldn’t have come up with otherwise. It’s essential for driving true innovation. However, typical AI tools accelerate a societal trend of avoiding friction because they are designed to be sycophantic. By constantly validating the user and agreeing with their ideas, AI strips away the healthy disagreement and compromise required to develop new, effective solutions.
Research demonstrates that while AI is highly effective at knowledge transfer, users actively resent it when it attempts to mimic human empathy (particularly in how it repeatedly asks reflective questions). Further, making AI act human ignores its actual core strength: rapidly sifting through vast amounts of complex data to provide immediate, actionable insights.
Strong ROI is demonstrated when the data shows actual behavioral shifts, such as managers successfully navigating difficult feedback conversations or a reduction in cross-functional silos. Traditional software metrics like login rates or time spent in-platform are flawed measures of success for AI coaching, as the goal is quick behavioral change, not creating an "attention economy."
To elicit real behavior change, AI coaching must be built on deep behavioral and relational data — such as personality assessments and HRIS data — from day one. This allows it to provide hyper-relevant insights the very first time it’s used. If an AI starts from "zero context," users won’t find value in the initial interactions and will quickly abandon the tool.
Delivering coaching in the "flow of work" means the AI doesn’t require a separate login or URL. Instead, it proactively meets employees where they already are, such as in Microsoft Teams or Slack. By pushing short, 30-second nudges containing the exact information an employee needs in that moment, the AI drives productivity without interrupting the workday.
Enact Meaningful Change With the Data You Already Have
The consensus is clear: AI coaching works best as an augmentation and democratization tool — extending coaching access at scale and handling routine development needs. What it isn’t is a replacement for human coaches in complex, emotionally sensitive or values-based scenarios.
Adoption curves can flatten without strong human integration If AI isn’t integrated into the tools employees already use, the likelihood they’ll continue to use it is low. The excitement is genuine, but organizations need smart deployments to take full advantage of AI’s potential to produce real results.
Cloverleaf can scale behavior change with a thoughtful approach to AI coaching. Learn more at cloverleaf.me.