One of my colleagues here at Improving is an agile coach. Not a developer. Not someone who works in code at all. But he does a lot of work that involves collecting information, facilitating processes, and helping people figure out what they need. When he found out I’d been using an AI agent to write user stories from stakeholder conversations, he asked if I could share it with him.
I sent the agent over. His follow-up message made clear he wasn’t quite sure what to do with it.
So we sat down together.
Building an Agent With Someone Who Doesn’t Code
I didn’t just hand him the agent and explain how it worked. I walked him through how I had built it, step by step. Not the technical mechanics, but the thinking: what problem was I trying to solve, what information did the agent need to work with, what output did I want it to produce.
And then we used that same agent to build a new one for him.
He runs a lot of informal support for people at the office who need help filing expense reports. He’d put together a slide deck over time with the most common questions and their answers. We used that material as the input for a new agent: one that could answer those questions for him, in his voice, with the context he’d already developed.
He has been refining and expanding that agent on his own ever since. Entirely without writing code.
The “AI Is for Developers” Assumption
When AI shows up in most conversations I’m part of, the examples skew heavily toward software. Generating code, writing tests, reviewing pull requests. That’s a real and useful application. But it’s one slice of what’s possible, and it’s the slice that already has a lot of attention.
What doesn’t get talked about as much is what AI can do for the people in every organization who are not engineers: the coordinators, the coaches, the operations people, the analysts, the support staff. The people who manage information, facilitate conversations, and keep things running.
Those jobs involve a lot of knowledge work that doesn’t require coding to automate or accelerate. Summarizing meeting notes. Drafting communications. Structuring feedback. Searching through documentation. Answering recurring questions consistently. AI handles all of that, and it doesn’t require the person using it to understand a single line of code.
What It Actually Takes
What it does require is clarity about the problem you’re trying to solve. That’s the actual skill.
My colleague wasn’t starting from “I want to learn about AI.” He was starting from “I spend a lot of time answering the same questions, and I’d like that to be easier.” That specificity is what made the session useful. We weren’t exploring the technology in the abstract. We were solving a specific thing for a specific person.
That’s the entry point for almost everyone who’s not technical: start with friction, not with features. Where is something taking more time than it should? Where are you doing something repeatedly that feels like it should be easier? That’s the problem to bring to AI.
The tool is a lot less intimidating when it’s solving something real.
Lunch and Learns That Actually Land
I’ve given a few informal sessions for teams at client sites that weren’t developer teams. Customer support, operations, back-office staff. The goal was never to teach them about AI in general. It was to show them, using examples from their actual day, how a specific kind of friction could be reduced.
Those sessions land differently from the generic AI overview talks. People leave not with a broad understanding of what AI can do, but with one concrete thing they can try tomorrow. That’s the right outcome.
The bigger picture is this: every organization that’s investing in AI at the engineering level also has dozens or hundreds of people who would benefit from AI support in their daily work and currently aren’t getting it. That’s not a technology gap. It’s a translation gap. Someone needs to sit with those people, ask what’s hard, and show them what’s possible at their level.
What I’m learning is that teaching someone to use AI well has very little to do with AI, and everything to do with helping them understand their own work clearly enough to know where help is needed.
We have great AI training and offerings at Improving.
Leave a Reply