The Sympathetic Technology Framework
Article
Snapshot
I did not come to AI very easily. All my friends are writers. I married a writer. And in the earliest days of LLMs, when they were landing in every single conversation in every single situation I was sure that these new tools were an abomination. The laundry list of lamentation was long. And then one day it changed. I ran head long into a reality that I could not just wish away. Not only were the tools not going anywhere, and sure, maybe they possessed some marginal utility, but a grim conclusion had been made clear to me: they could be used against us. That changed everything.
There’s a version of AI adoption being sold right now that goes something like this: replace the expensive human parts with cheaper automated parts, watch productivity go up, declare victory.
That’s not what I do. And I think it’s going to cost a lot of organizations dearly before they figure out why it doesn’t work.
The Sympathetic Technology Framework starts from a different premise. Tools should be in service of people — not the other way around. That sounds obvious. It isn’t. Because the pressure right now, in almost every organization I work with, runs in exactly the opposite direction. The tool arrives, the vendor makes promises, leadership gets excited, and suddenly the humans are being asked to adapt themselves to fit the workflow the software prefers. The org reorganizes around the tool. The tool does not reorganize around the org.
Sympathetic technology means the opposite. It means deploying AI in ways that extend what your people are already good at — using it to handle the work that drains energy without creating value: the summarizing, the scheduling, the formatting, the first draft of the thing nobody wanted to write. And protecting the work that actually requires a human: the judgment call, the relationship, the read of the room, the decision that has values baked into it.
The word sympathetic is doing real work here. In physics, sympathetic resonance is what happens when one object starts vibrating because another nearby object is vibrating at the right frequency. The second object doesn’t get replaced. It gets amplified. That’s the model. AI at its best is a resonance — it picks up what you’re already doing and makes it louder, faster, further-reaching. It does not replace the original signal.
This matters especially for the kinds of organizations I work with — healthcare associations, cultural nonprofits, mission-driven businesses — because their work is irreducibly human. Their value isn’t processing speed or information volume. It’s trust, relationships, expertise, and judgment built over years. I came up in the literary world, spent two decades in professional associations and nonprofits, and what I saw consistently was that the organizations people actually love are the ones that understood this. You cannot automate your way to that. You can, however, use AI to free up the people who have it so they can do more of it.
The failure mode I see most often isn’t organizations that refuse to adopt AI. It’s organizations that adopt it without a framework — that hand the tools to people without a clear sense of what they’re for, what they’re not for, and where human judgment isn’t up for negotiation. The result is inconsistent output, eroded voice, and a creeping sense that something important has gone missing. It usually has.
Sympathetic technology isn’t a cautious position. It’s an ambitious one. It says: use every tool available, use them well, and measure success by whether your people are doing better, more meaningful work — not whether you’ve managed to need fewer of them.
I built this framework because I needed it myself. I still do.
