If you’re reading this, I’m guessing you’ve already heard that AI is going to change everything (spoiler alert: already has).

Earlier this week, I read through Matt Shumer’s essay “Something Big Is Happening”, and I encourage you to read it as well. In it, Shumer, CEO and Co-Founder at OthersideAI, lays out a very convincing case for AI fundamentally changing the lives and careers of everyone who works at a computer. Not just folks in the tech industry, but doctors, lawyers, accountants, financial analysts, customer service reps, etc. As Shumer puts it: “I think the honest answer is that nothing that can be done on a computer is safe in the medium term.”

He’s not wrong. I agree that if you do any amount of work at a computer, whether it’s writing grants, analyzing data, communicating with stakeholders, and certainly even managing AI agents, you need to be engaging with AI tools now. Not eventually. Now. The cost of waiting to understand what these tools can do is growing every day.

While Shumer’s framing does an excellent job of jolting you awake (seriously, give his essay a read), and seems to come from a place of care and empathy for his readers, no matter how many times you say “Don’t panic!”, inevitably, people will panic. So how can we retain agency over our careers and lives in the face of this monumental shift?

The Treadmill We’ve Been On Before

Here’s a pattern worth paying attention to: every major productivity tool of the last forty years has promised to save us time. Computers, emails, smartphones, even Slack (note: if you actually want to save time with Slack, just put it in “Productivity Mode” by hitting Cmd+Q on a Mac or Alt+F4 on a PC…). Ultimately, these tools didn’t “save” time, they just shifted that time to other tasks, and it was the employees and workers who felt the squeeze. Remy Reya has a name for this cycle: the Impact Treadmill.

In a presentation titled “Using AI to Make Nonprofit Work More Human”, Remy highlights the tension between “time saved” and “finding more work to fill that time”. At his organization, Compass Pro Bono, staff saved 10-12 hours per week using AI tools – a remarkable gain. But when surveyed about what happened to that time, 33% said “nothing specific. They shared that time just gets absorbed into the day.”

AI is the latest entrant in a long line of technologies that promise to give us our time back. The question isn’t whether AI saves time, but rather what happens to that time once it’s saved.

What You Already Know Is Your Advantage

Here’s what I find exciting about this moment, especially for people leading mission-driven organizations: you already know something that most of the tech industry is still figuring out. You know who you serve and why.

Nonprofits, foundations, civic tech organizations, and social impact startups have spent years understanding their communities, building trust, and doing the unglamorous work of figuring out what people actually need. That knowledge doesn’t get automated, and it doesn’t show up in a prompt. And it’s exactly the kind of judgment that AI doesn’t replace.

The challenge facing these organizations now is translating that mission and all of that hard-won knowledge about their communities, their programs, and their impact, into the kind of digital tools and workflows that can grow what really matters. The AI conversation has been dominated by people like Shumer who live in the tech world, and their framing around speed, scale, and replacement, doesn’t map to the realities of everyone’s lives.

For those who think of AI as a powerful tool that can do a lot of good, I propose a different starting point.

A Way Out: What Job Are You Hiring AI to Do?

Clayton Christensen, probably best known for coining the term “disruptive innovation,” spent years studying why people buy what they buy. His insight was simple: people don’t actually “buy” products. Instead, they “hire” them to do a job. And often that job is a very specific and precise job. Check out one of his more famous (and for the green-screen aficionados, maybe infamous?) examples of explaining his Jobs To Be Done framework in this short video.

I think this framework provides a great off-ramp from the anxiety and fear felt by a lot of people working at computers when it comes to AI. Instead of asking “what will AI replace?”, a question that’s paralyzing at the individual level and leads to the dreaded treadmill at the organizational level, ask: what job is concretely in front of me right now? And can I hire AI to do it?

If you’re an executive director drowning in grant reporting, the job isn’t “use AI.” The job is “help me synthesize our program data into a narrative that funders can understand so I can get back to talking to the people we serve.” If you’re a program manager buried in survey data, the job isn’t “analyze everything.” It’s “help me find the three patterns in this data that will shape our next quarter so I can spend my afternoon in the community.”

See the difference? The JTBD lens keeps you grounded in your mission and your context. It turns AI from an abstract threat into a concrete tool you can evaluate:

• Did it do the job I hired it for?
• Did I use the time it freed up for something that matters?

If not, try a different job. If so, try another one!

The people and organizations who will end up using AI well (and I’d argue who are already using it well) aren’t the ones moving fastest. They’re the ones who slow down enough to ask what they actually need. They’re the ones who lean hard into the human work that no model can do for them.

Try It

If you’re new to AI tools, or if you’ve been circling it warily, here’s my suggestion: don’t start with the big questions. As much as the marketers show you ads with simple short prompts that lead to grandiose answers, it behooves you to not be taken by the magic of it, and to treat it as a tool that has right and wrong ways to use it. Don’t start with “what does AI mean for my organization?” Start with one job. Something concrete that’s on your plate this week. Give it to Claude or ChatGPT and see what happens. Pay attention to what comes back, and more importantly, pay attention to what you do with the time you get back.

That’s the real test. Not whether AI can do the work. But whether you use what it gives you to work that matters.