Kimberly Goes to Business School #3: When the Tool Isn’t the System

Kimberly Goes to Business School is a numbered series within my Business Different column. It chronicles Biscochito’s field study with Harvard Business School and Tulane University.

A few weeks into our study with Harvard Business School and Tulane University, we hit an operational tension.

Caregivers were successfully coordinating shift coverage, but no one was using RobinaAI. Not once. Not even opened.

The shift from centralized scheduling to collaborative coverage had already been made. It wasn’t perfect, but it worked. Caregivers were responsible for requesting coverage, responding to each other, and ensuring shifts were filled. The system was nearly intuitive: open WhatsApp, open your client-specific group, type your request. Nina Martinez and Roberta Payne supported the edge cases so client coverage stayed uninterrupted.

It was working, though starting to show strain. Conversations occasionally overlapped, caregivers sometimes posted requests in the wrong group, and needs were missed. More concerning, the AI tool we were meant to be testing the impact of wasn’t being used at all.

I flew to New Orleans to meet with the research team at Tulane University, excited about the work, nervous about the engagement gap, and thrilled that David Byrne was headlining Jazz Fest on Sunday and we all had tickets.

Before I left, Maxine Swisa and I called every caregiver. No coaching, no reminders to try the tool, just curiosity. We split the calls, she spoke with the RobinaAI group, I called the non-AI group and we listened to each other in our shared workspace.

The patterns were immediate.

Everyone feels a little, or a lot, off when they ask for help and no one responds. A caregiver who’d been part of the organization for over a year left during this period. She likely would have moved on regardless, but she named the shift to more collaborative work as a reason. We noted it.

And no one, not one person, was going to take an extra step on their phone before going into WhatsApp.

The problem was workflow. Caregivers were willing to change how they coordinated coverage, they had adopted the WhatsApp approach without resistance. But RobinaAI wasn’t part of that workflow. It was a step outside it. The WhatsApp track required one action. RobinaAI required two. That extra step, sitting just outside the actual work, was enough to eliminate engagement entirely.

I walked the research team through all of it: the call data, the patterns, the single-step problem. I had a list of ideas: ways to encourage usage and structure adoption.

Jake Krupka, PhD, part of the Tulane research team, listened, then offered something simpler.

The tech had evolved. RobinaAI no longer needed to be a separate step, it could be embedded directly into the WhatsApp workflow. One step. The exact amount of change caregivers were comfortable with.

The list of ideas stayed in my notebook.

On Sunday, the researchers and I went to Jazz Fest. David Byrne, at 73, gave a masterful performance, completely in tune with his band and dancers, every movement flowing, the entire stage connected. Watching him, I found myself thinking: I can probably keep doing good work for another twenty years, too.

I had expected to spend the next phase of this study driving adoption. Instead, we’re integrating.

Caregivers were already coordinating effectively. Any supporting tool needs to exist inside that behavior, not adjacent to it. A tool that sits outside the system will go unopened.

The constraint isn’t motivation.

It’s friction.

Often, it’s one extra step.

Now we integrate and find out whether the design carries the weight.

This column ran in the Santa Fe New Mexican on May 5, 2026, under Business Different.