What Leaders Tell Themselves About AI : Some Advice and Tips
I’ll start by acknowledging the mixed feelings everyone has about AI - what we know about it and what we don’t know about it. Its impact on brains and the environment. What I know is many of your employees are using it and without guidelines it is posing a business risk. Here is more about our approach and things we have learned.
Last month, I gave our team 30 minutes and a simple prompt: tell me what’s hard in your work right now. Then, we split into two groups, each paired with Claude, and got to work. Half an hour later, we had three tools we’d never built before — an RFP screening matrix, an implementation assessment for onboarding new clients, and an environmental scan search tool we’ve already used twice since.
All of that in thirty minutes.
I share this because I know what many of you are thinking. You’re too busy to add one more thing. You’re not sure it’s safe. You’re not convinced it’s worth it. Or maybe you’re quietly hoping AI will crest like every other tech wave and you can wait it out. Or you have team members that just don’t want to use it. I hear you. Here are some of my thoughts…
First: How We Actually Use AI at The Spark Mill
Before we get to the myths, here’s the framework we’ve landed on — not because it’s sophisticated, but because it’s honest about what we have discovered AI is actually good for and it fits with our values.
The Human Sandwich. You do the thinking, AI does a draft or generates options in the middle, and you finish it. You start and you end. AI lives in between. This keeps your judgment in charge while letting AI do the heavy lifting on structure, synthesis, or first drafts. The output is only as good as the human on both ends — which means your expertise still matters enormously.
The Brainstorm Partner. When you are stuck on a problem or need to see it differently, ask AI to push back, generate alternatives, or play devil’s advocate. This is especially useful for leaders who are isolated in their roles and don’t always have a peer to think alongside.
Ask Me What I’m Missing. This one is underused and underrated. Paste in what you’ve got — a memo, a meeting agenda, a grant narrative — and ask: “What have I not thought about? What am I missing?” The AI will surface gaps your brain has tunnel-visioned past. We don’t use it as a replacement for good thinking, but it is a useful set of second eyes.
Here’s Some Push Back I Frequently Hear
Myth #1: “I Can Just Avoid It”
This is the most common thing I hear from senior leaders, and I understand the impulse. You’re already overwhelmed. Learning a new tool feels like a luxury, not a strategy. So you figure you’ll let it play out.
Here’s the problem: it’s already playing out inside your organization. Gallup’s Q2 2026 survey of more than 22,000 employed US adults found that 52% of workers now use AI in their role — up from just 21% in 2023. That’s more than doubled in three years. Your team isn’t waiting for a policy. They’re experimenting right now, in their browser tabs, on their own terms.
For those of us in the nonprofit world, the gap is even more striking. According to TechSoup and Tapp Network, 85.6% of nonprofits are already exploring AI tools — but 76% have no AI policy at all. That combination of broad experimentation and minimal guidance is exactly where things go sideways.
This is what researchers call “shadow AI” and it’s not a fringe behavior. It’s what happens when tools are powerful, free, and accessible, and organizational guidance doesn’t exist yet. A 2026 survey found that two-thirds of office professionals have used AI tools at work despite believing they were not permitted under company policy and many are feeding sensitive business data into public models in the process.
Not having an AI policy isn’t a neutral stance. It’s a liability. There are more controls in place if you have an enterprise plan versus a million people dabbling in their personal accounts.
Myth #2: “AI Is Going to Replace My Staff”
The dismissive “AI won’t replace people” response doesn’t help us right now. The reality is more nuanced and ultimately, more hopeful.
The fear is real and widespread. A 2026 Pew survey found that 52% of U.S. workers worry about the future impact of AI on their jobs. And 18% say it’s likely their job will be eliminated within five years — up from 15% just a year ago. Your staff probably have this fear. Naming it is better leadership than pretending it doesn’t exist.
But here’s what the evidence actually shows: roles that require unpredictable decision-making, human connection, and ethical judgment are the hardest to automate — and that’s exactly the kind of work most of your team does. AI can draft a grant narrative; it cannot build the relationship with the program officer. It can screen an RFP; it cannot decide whether the work aligns with your mission and values. It can summarize a meeting; it cannot hold the room.
Myth #3: “It Takes More Energy Than It’s Worth”
I’ve heard this from a lot of smart, busy leaders: using AI feels like more work than just doing the thing. That is true when it is used poorly.
Researchers have a name for it now: “botsitting.” The Work AI Index 2026, which surveyed 6,000 workers across the US, UK, and Australia, found that the average worker burns 6.4 hours a week feeding AI missing context, checking outputs, debugging mistakes, and cleaning up answers that are confident but wrong. That’s nearly a full working day, every week, spent managing a tool rather than using it.
But here’s what they also found: the organizations where workers aren’t losing time to AI have one thing in common. They’ve made the approved path the easiest path. Clear guidance on which tools to use, for what, and how — and humans staying in charge of the output. That’s the Human Sandwich. The people who find AI exhausting are the ones using it as a vending machine — put in a prompt, get out a mediocre result, feel disappointed, spend an hour fixing it. The people who find it genuinely useful pick a specific problem, use a clear approach, and stay in the loop from start to finish.
The time savings are real when you work it that way. The Federal Reserve Bank of St. Louis found that AI users save an average of 2.2 hours per week — and among daily users, one in three saves four or more hours. That’s not magic. That’s the compounding effect of better first drafts, faster research, and less time staring at a blank page.
And remember: AI is already embedded in the tools you use every day. Microsoft 365 Copilot is baked into Word, Outlook, and Teams. Google’s AI is woven into Search and Docs. You are already using AI — you’re just not always the one directing it. The possibility is in using a specific tool, for a specific reason, with a clear process.
Where to Start
If you’re a senior leader or a manager who has been quietly skeptical, my advice is this: you don’t need to become an AI expert. You need to become an AI-aware leader.
That means a few things practically:
Get curious before you get prescriptive. Try one of the three moves above on a real problem before you write a policy or issue a directive. Your credibility with your team depends on your firsthand experience.
Have the conversation. Ask your team what they’re already using. You might be surprised. And that conversation is the beginning of the policy you actually need.
Build the policy from practice, not before it. A good AI policy doesn’t come from a template. It comes from paying attention to where your work actually touches AI and making clear decisions about what’s okay.
At The Spark Mill, we’ve put together a working internal policy and we are happy to share it with you as a starting point. It’s not finished — no policy in this space is. But it’s honest, it’s grounded in how we actually work, and it’s a better foundation than silence.