The Text That Changed How I Think About AI Communication.

I Was Reviewing Deployment Logs on a Tuesday. This Exchange Stopped Me.

I was reviewing deployment logs on a Tuesday afternoon, something I do most weeks just to see what's actually moving through our systems.

One exchange stopped me.

It was late at night. Someone had reached out to a human services organization we work with, a nonprofit that supports families in their community. The message was simple on the surface. A question about services they offered. Eligibility requirements. The kind of thing that ends up in an FAQ doc.

But the way it was written, you could feel something underneath. Not panic, exactly. Just the kind of quiet uncertainty that comes when someone is trying to figure out if there's help available for them and they're not sure they're going to like the answer.

The AI responded. It answered the question clearly. Explained the services, walked through what they'd need to qualify, and pointed them to the next step. Warm, accurate, straightforward.

But here's what actually moved me: the person wrote back.

Not to ask another question. Not to schedule something. They wrote back something that made clear they felt received. Not understood by a person necessarily, but met by a system that had been designed with enough care that they could tell someone had thought about them.

I sat there for a minute just reading that exchange.

The Bar Is Not Whether the AI Answered Correctly

This was the moment I understood something that I now think about every single time we deploy: the bar is not "did the AI answer correctly?" The bar is "did the person on the other end feel like they mattered?"

Because people can tell the difference between a system that's designed to care and one that's just designed to process them.

A templated response with mail-merge feels like mail-merge. A genuinely considered message, even if it's generated by AI, feels like someone thought about you.

That person didn't know it was AI responding. They just knew that when they sent a message late at night, trying to figure out if help was available for their family, something took their question seriously and gave them an answer that felt like it was written for them.

What That Exchange Recalibrated for Me

That exchange recalibrated something for me.

I oversee deployments across nonprofits, healthcare organizations, and universities. My job is partly to make sure the technology works. But what that exchange made real is that the technology is just infrastructure. The actual work, the part that matters, is whether the system treats people like people.

I started asking different questions after that.

Not: "Is the accuracy at 97%?"

But: "Would someone feel rushed by this response? Would they feel like we care about their situation? Would they know we heard them?"

It sounds soft when I write it down. But it's the difference between a system that's technically excellent and one that actually works.

The People on the Other Side of These Decisions

Here's what gets me about that particular exchange: that person was real. Not test data. Not a benchmark. A real person, at a real moment, trying to figure out if there was somewhere to turn.

I think about that a lot when the work gets abstract. When we're deep in architecture decisions or optimization conversations. It's easy to lose sight of the fact that on the other side of these decisions are real people in real moments of need.

A family trying to find out if they qualify for services. A nonprofit volunteer coordinator getting a reminder about their shift. A student figuring out where to go for support. These aren't edge cases. These are the actual job.

And the weight of knowing that, really knowing it, not just intellectually but in your gut, changes how you build.‍

‍What We Are Actually Building

The thing that stayed with me most wasn't even the exchange itself. It was what it taught me about what we're actually building.

We're not building automation. We're not even building efficiency. We're building trust infrastructure. Systems that let people feel like they matter to the organizations that serve them.

That's different. That's harder. And it's worth doing carefully.

When was the last time an automated system made you feel heard? What made the difference?

‍

More guides on this topic

• Your Website Isn't Serving You As Well As You Think
• Your Families Check Phones 96 Times. You Send Emails.
• Your Staff Won't Resist AI. They'll Resist the Way You Roll It Out.
• AI Chatbot vs Live Agent for SMS: When to Use Each