Before we begin, there’s something important I want to share with you: next month at The Big Game, our flagship event, we’ll make our biggest announcement yet. I’m excited to share what we’ve been building, and more importantly, what it means for how companies orchestrate the hybrid frontline to drive real outcomes. I hope you’ll be there with us.
You can either join in-person at London’s Tottenham Hotspur Stadium, or you can watch the virtual broadcast from anywhere. Spots are limited. Register here.
The first year I coached volleyball, I trained all twelve kids on my team the same way: fundamentals, repetition, discipline. It worked; we got good results.
Then I read Power Volleyball by Arie Selinger, coach of the US women’s national team.
He broke every movement down to the individual athlete, even the circular arm swing, and matched each technique to a player's height and joints. In my second year, I coached each kid based on their body. Their progress was amazing.
Thirty years later, I watch team leaders make my first-year mistake, with a twist.
Watch any team leader for a month, and a pattern shows up: the same few people get most of the coaching, on the same few things, again and again. Not the people who need it most, but the people the manager already has an eye on.
We call it judgment, but it’s simply habit. This isn’t just my opinion: at Centrical, we ran the numbers on 834,000 real coaching interactions. Keep reading to see what we discovered.
It’s easy to blame the manager, but even the best leaders tend to fall into this pattern. Give someone a team of 15 and a full day of work, and coaching narrows to whoever is top of mind: the new hire, the one who’s struggling, the person who just took a brutal call.
So what can a manager actually see from where they stand?
I recently spoke about this with Tal Eden, our Head of Data and Analytics and a clinical psychologist by training. His answer: this is how human cognition works.
It comes down to two limits.
First, human bias:
“Managers tend to reach for the interventions they know best and to focus on the employees they already pay attention to, so coaching becomes repetitive and some people receive far more of it than others.”
The second limit is limited working memory:
“A manager can track only a handful of performance signals at a time, so most of the available data never factors into the decision.”
So, how can AI help?
“AI is not subject to either limit. It reviews every relevant metric for every individual, pinpoints the specific behavior each person needs to work on, and selects the intervention with the strongest evidence behind it, unaffected by habit or by how much a person can hold in mind at once.
Its role is to augment the manager, not to replace them. The recommendation still reaches the employee through the manager, which preserves the relationship that makes coaching work: trust, accountability, and the credibility of a familiar figure. AI supplies what the manager cannot see or process, and the manager supplies the relationship in which behavior actually changes.”
Two blind spots, both hardwired, neither fixed by experience.
Picture what sits in front of that manager: 15 people, each producing quality scores, KPI trends, learning progress, and a dozen more signals every week.
The data to coach every person precisely already exists. What’s missing is a human who can hold all of it, rank it, and act on it. So managers fall back on the two or three names they can carry in their head, and the rest of the team remains invisible. Not because the manager doesn't care, but because no one can see that much at once.
Many operations run a regular coaching cadence. But structured cycles only guarantee that coaching happens, not that it reaches the right person on the right KPI. Managers juggle many competing priorities, and they often don’t have the bandwidth to know who needs support and on what.
This is the question every frontline leader is now trying to answer: what do you hand to AI, and what has to stay human?
If you automate the relationship, coaching stops working. If you leave the targeting to memory, most of your team never gets the attention they need.
We went looking for where the line actually falls.
We analyzed 834,000 coaching interactions in the Centrical platform from a top-5 banking and financial services organization and a global hospitality brand.
AI-guided coaching improved the targeted metric 59.1% of the time, compared with 48.9% without it, a gap of about 10 points. The difference was the precision of the coaching. And it wasn’t one metric moving: 19 of the 20 most-coached KPIs improved.
The full breakdown is in our latest Centrical Labs analysis:
Here’s what AI is doing in that gap: it reads every relevant signal for every person and points the manager to who needs help, on which metric, with the intervention that has the strongest evidence behind it. It doesn't have the conversation.
AI doesn’t build the trust. The recommendation still reaches the employee through their own manager.
So the choice was never AI or the human coach. AI makes the manager’s coaching more precise.
If you lead a frontline team, that reframes what to optimize:
Use AI to handle the targeting. Deciding who to coach and on what is a data problem that AI can help you solve.
Protect the human conversation. It works because a person your team trusts is the one delivering it.
This is the split we built our AI Coaching Assistant around. AI targets the right coaching opportunities, the manager does the coaching, and the recommendation always travels through a human. Our data shows this positively impacts the team’s performance.
And it falls within a much larger shift happening right now: how to orchestrate a frontline where AI and people each do what they do best.
We’ll get into it at The Big Game on October 15, where I’ll be sharing our biggest announcement since I started Centrical. Join us at Tottenham Hotspur Stadium or online from anywhere to see how leading CX and operations teams are orchestrating the hybrid frontline today.