Teaching a Machine to Protect What Makes Us Human

Maria, one of your Care Professionals, has been quietly disengaging for two months. Nobody flagged it. Her numbers are fine, her clients love her, she shows up on time. The kind of quiet drift that never trips an alarm until the day someone hands in their notice.

What caught it wasn’t a gut feeling. It was a pattern your team now tracks: a small dip in her engagement responses, a missed recognition milestone, a stretch of weeks without a single one-on-one logged. A system flagged it. A real human on your team had the meaningful conversation.

That’s the whole argument for this newsletter in one story.

Turnover – A Leadership Problem

I’ve said this for years and I’ll keep saying it: when good people leave, it’s rarely because they stopped caring. It’s actually often because leadership ran out of time to show they cared back.

You know the ARCC pillars by now if you’ve been reading my newsletters the last few months. Autonomy. Relatedness. Capability. Confidence, the outcome all three produce when they’re working together. None of that is complicated in theory. What’s hard is doing it consistently, for every care professional and office team member, every week, while you’re also running payroll, fielding a family complaint, and putting out whatever fire showed up before lunch.

Turnover is often due to the fact the leader ran out of capacity to care back.

And capacity is exactly what AI is good at giving back.

Where AI Actually Helps

Think about the layers of the KEEP Culture Model. Core4 Alignment. ARCC of Retention. The KEEP Communications Process, Know, Envision, Engineer, Purpose-driven action. Every layer depends on someone noticing something, and then someone acting on it.

AI is very good at the noticing. It can track engagement signals across a hundred caregivers at once, something no human leader can hold in their head. It can flag when Autonomy is slipping, a caregiver whose schedule has become rigid instead of self-directed. It can catch when Relatedness is thinning, a team member who hasn’t had a real conversation with their supervisor in three weeks. It can remind a leader that Capability needs reinforcing, that someone’s due for a skill check-in before their confidence quietly erodes.

Used this way, AI becomes the early-warning system your Communications Process needs to actually run on schedule instead of whenever a leader remembers to get to it.

That’s the Know part of the process. It’s also the part I got wrong for years, hoping intuition and good intentions would catch what a system now catches automatically.

Where AI Cannot Go

Here’s where AI cannot go: AI can tell you Maria is drifting. It cannot sit across from Maria.

It can’t hear the hesitation in her voice when she says “I’m fine.” It can’t read the room when a caregiver says the right words but means something else. It can’t offer the kind of Envision what’s possible conversation that comes from a leader who genuinely knows someone’s Guiding Truths, and can connect it back to why this work matters to them specifically.

The Engineer the way forward and Purpose-driven action steps of KEEP Communications are fundamentally human work. They require presence. They require a leader, such as a Care Coordinator, willing to sit in an uncomfortable silence instead of filling it. No tool replaces that, and I’d be lying to you if I suggested otherwise.

This is where I’ve watched founders get it backwards. They install a great retention tool, see the dashboard light up with insights, and mistake the insight for the intervention. The data told you where to look. It didn’t have the conversation for you.

If your team starts treating a flag in a dashboard as the finish line instead of the starting line, you haven’t built a KEEP Culture. You’ve built a really well-informed one that still isn’t keeping anybody.

What This Actually Looks Like

At Riverside Home Care, we’re building the capability that addresses the need for AI-assisted engagement tracking the way I’d treat a good scout in football. It watches film nobody has time to watch. It tells the coach where to look. But it never calls the play.

Practically, that means a system flags the signal, Autonomy dropping, Relatedness thinning, whatever it is, and a leader is expected to have a real, in-person or at least real-time conversation within days, not weeks. The tool creates urgency and accountability around the noticing. The human owns everything after that.

We’ve also learned the hard way that this only works with consistent integration. I’ve told you before about the two years it took to transform communication culture in a family business I worked with, and how quickly it regressed when new hires weren’t trained into it. The same is true here. A brand-new AI tool doesn’t fix a culture that hasn’t committed to the human follow-through. It just gives you faster, better data about the same neglect.

The Test I'd Give You

Before you adopt any AI tool aimed at culture or retention, ask one question: does this [Ai tool] get a leader into a real conversation faster, or does it let a leader believe the conversation already happened?

If it’s the first, it’s protecting the human element. If it’s the second, it’s quietly replacing it, no matter how good the dashboard looks.

If you're building a culture that's meant to outlast you, and you want to see how we're thinking about AI, retention, and the exit that follows both done well, visit www.rivhc.com/founders.

P.S. The best retention tool I know of still doesn’t run on electricity. It runs on a leader who shows up. AI just makes sure they show up for the right person, at the right time.