TL;DR: If email and a smartphone are about as far as you’ve gone with technology in twenty years, treating AI as a tech thing for tech people makes sense. It’s the wrong read for this wave, because the interface is a conversation: you type a sentence and get an answer, with no setup, configuration or learning curve. You can see what that looks like on a real project in how I use AI on a book.
I’ve got a lot of respect for professionals who skipped the last few waves of tech hype.
That skepticism saved them a pile of wasted afternoons. This wave works differently, because you talk to it in plain English, and I’d hate to see good people sit it out on a rule that only fit the old waves. Minimal adoption doesn’t require becoming a tech person, and there are still jobs where staying out is the right call.
Who this article is for
You’re an experienced professional with real expertise in your field. The expertise was built without much help from technology beyond the basics. You use email, a phone, and whatever software your industry forced you to learn. The world of apps, cloud tools, productivity systems, and tech enthusiasm isn’t your world and never has been.
You’ve watched colleagues get excited about successive waves of technology that mostly faded, and your skepticism has saved you from a lot of wasted time. The “I don’t really use computers” stance has worked.
The argument isn’t that you need to become a different kind of professional.Share on X
Nobody’s asking you to change who you are or become a different kind of professional. There’s one case where the skepticism that’s served you well gives the wrong answer, and I think it deserves a hard look. The cost of getting it wrong adds up slowly, so you may not see the bill for years.
Why is this AI wave different for non-technical professionals?
Past technology waves required you to learn something to participate. Personal computers required you to learn an interface and a set of file management concepts. The web required you to learn how browsers and links worked. Smartphones required you to learn an app model and gestures. Social media required you to learn each platform’s specific conventions. Each one had a learning curve. The curve was steep enough that staying out was a reasonable decision for plenty of older professionals who didn’t want to spend the time.
AI doesn’t have that learning curve, and holdouts often miss it. You type a sentence in plain English, the way you’d write an email or ask a colleague a question, and the system answers in plain English. There’s no syntax to learn and no apps to configure. If you can write a clear email, you’ve already got the only skill it asks for.
The conversational interface is the closest thing in computing history to “no curve at all.” That changes the calculation for the non-technical holdout, in a way the holdout should notice.
What does minimal AI adoption look like?
One real task a week handed to AI, checked carefully, and kept only if it saved time without costing accuracy.
From Conversations With Influencers
Abbie Richie brings patient tech support to senior living communities, starting with five free workshops in one day and reaching thirty-five states after pivoting to Zoom. She described the approach on Conversations With Influencers.
Minimal adoption doesn’t mean becoming someone who follows AI news, tries new models, or talks about prompts.
Pick one task you do every week that involves writing, summarizing, or organizing. Open a single AI chat tool. Use it on that one task. That’s it. The task might be a difficult email you’ve been putting off. A long document to summarize. Meeting notes to organize. A memo to draft from a few rough points.
You type what you want in plain English. The result comes back. You decide whether to use it. If yes, you keep what’s useful and adjust what’s not. If no, you ignore it and continue as before. The total commitment is the time it takes to try once, and the upside is that you may discover one weekly task that becomes considerably easier. That discovery is the entry point. The skepticism that served you well in past waves can remain intact, because nothing about this experiment commits you to anything broader.
I’d start with the task you dread most, like the email you’ve rewritten four times in your head or the report nobody reads but somebody has to write. If the tool takes some of the misery out of that one job, you’ve learned something real about it. If it doesn’t, you’ve lost an hour and you can go back to ignoring it.
When is staying out of AI still right?
Some categories of professional work don’t benefit from AI assistance, and the holdouts in those categories are right to stay out.
Surgical practice. Live performance. Therapeutic presence with patients. Master craft trades where the value is the hand and the eye. Direct teaching where the student needs the teacher in the room. Negotiation that depends on reading a room of specific humans. Diagnostic work that depends on physical examination.
In all of those categories, the working day doesn’t involve much writing, summarizing, or organizing of text, and the AI tools have nothing meaningful to offer.
If your work is in one of those categories and the work hours are spent doing the irreplaceable thing, then your “this isn’t for me” stance is correct.
The skepticism is doing its job. The case for engagement applies to professionals whose work does include text generation, document handling, research, or communication at real volume. A real share of those tasks could happen faster with the tools. Most knowledge work falls in this category. Pure craft work and pure presence work usually don’t.
I’d say the same to anyone in those fields who feels pushed. A surgeon doesn’t need a chatbot in the operating room, and any vendor who says otherwise is selling something. Nobody should feel behind for skipping a tool that has nothing to do with their work.
The “I’ll have my assistant do it” non-solution
A common response from senior professionals is to delegate AI use to an assistant or younger staff member. That delegation is half a solution. The assistant can do the mechanical work of running the tool, and the assistant’s familiarity will save the principal real time. That other half can’t be delegated. Judging when the tool’s output is good and when it’s not depends on the principal’s own expertise.
An assistant can’t tell whether a draft captures the principal’s voice, includes the right specific facts, or makes the argument the way the principal would make it.
I care about this one because voice is my whole job. When I write a client’s book, the words have to sound like the client, and only the client can finally say whether they do. An assistant running a chatbot can’t make that call for an executive any more than I could approve a chapter on a client’s behalf.
The pattern that works is the principal develops a basic familiarity with the tool through a few sessions, then delegates the routine work while staying involved in the judgment.
The familiarity doesn’t need to be deep. A few hours of direct use are enough to develop the working sense of what the tool does well and what it does poorly. After that, the assistant runs the mechanics and the principal supervises the output.
The principal who skips the familiarity step and delegates entirely tends to get burned, because they can’t tell when the assistant’s AI-assisted work has drifted in ways the principal wouldn’t have approved. Drift cost is sometimes significant. The hallucination survival guide covers the cases where the principal needs to be the one verifying. The principal can’t delegate the verification because the expertise it tests is the principal’s.
The cumulative effect, said gently
Nobody needs to act this month. The benefit builds over years. A professional who tries the tool once this year and builds the habit over the following months will be ahead of the one who tries it for the first time three years from now, under deadline pressure, with a client waiting. That second professional is the one I worry about.
The reason the timescale matters is that the comfortable late entry the skeptic is counting on may not exist. Past technology waves had natural late-entry points where the technology stabilized and a tutorial-driven catch-up was possible. AI is changing too fast for that stabilization to arrive on a predictable schedule, and the working judgment that matters is built through use instead of through reading.
The professional who has used the tool a few hours a month for two years will have judgment that the professional who reads about it for two years won’t. The asymmetry grows slowly, and the cost of waiting is paid in the experiential gap instead of in any single moment of disadvantage. A piece on practical AI adoption covers the working version of this for non-technical professionals.
What single AI experiment should a non-technical professional try?
If you’ve read this far and your skepticism is still intact, you owe yourself the same experiment recommended for any thoughtful doubter. One task, one tool, two hours. Pick something you do every week. Try one AI chat tool on that task once. Notice what happened. The cost is two hours. The upside is real information about whether the tool helps your work, the question only your own experience can answer.
If the experiment confirms your skepticism, you’ve justified the holdout position and can stop reading articles about AI without guilt. If it surprises you, you’ve found something useful. Either way, you’re working from your own judgment, and I respect that far more than skepticism inherited from the last three tech fads.
