If You Were an AI, I'd Think You Were Terrible

"Dude, if you were an AI and I was a normal person, I would think AI is terrible."

I sent that text to one of the smartest people I know today. He laughed because he knew exactly what I meant.

Here's what happened: About a week ago, he told me he was going to do something specific for me. Then he did something similar and sent it via email, but never actually did the thing he promised. So today I texted: "Hey, can you send me the thing?"

"I sent it to you via email."

"No, I meant this thing."

"Yeah, I sent that thing to you via email."

"Unless I'm mistaken, no you didn't. I just looked. I need this other version, which you promised here (link to text from a week ago)."

"Oh no, LOL. I totally overlooked that part. My bad."

You know what I did? I laughed. Because I know he's brilliant and well-intentioned. He just misunderstood. No big deal.

But if AI responded exactly that way—missing what I asked for, insisting it did the thing, then admitting it overlooked something—I'd be furious. Most people would. We'd conclude AI is either incompetent or broken.

That's the double standard that's killing your AI collaboration.

When I interact with a human teammate, I start from a fundamental belief that they're thoughtful, capable, and trying their best. If they miss something, I give them the benefit of the doubt. I clarify. We iterate.

AI gets none of that grace. And that reveals everything about how we're actually orienting to it.

Here's What Actually Needs to Shift

I'm giving four keynotes this week (yes, four—it's been a week). Everyone keeps asking the same question: "Okay, but how do I actually treat AI like a teammate?"

Two mindset shifts that work together:

Give AI the benefit of the doubt. When it doesn't nail something, assume it misunderstood—not that it can't. That's what I did with my whip-smart collaborator because that’s you do with teammates you trust. You don't immediately conclude they're useless. You clarify. You give them another shot. You provide more context.

Raise your expectations. Because here's the bizarre thing about AI—it performs to your expectations. (I know this sounds woo-woo, but stay with me.) If you don't think it can do something, it probably won't. Not because it can't, but because your low expectations produce lazy prompts. Lazy prompts produce mediocre outputs. And then you go, "See? I knew AI couldn't do that."

Self-fulfilling prophecy.

The truth? You failed to expect enough to prompt well enough to give it a fair shot.

Five Things You Can Do Tomorrow

In the spirit of hyper-practicality, here's what I've been telling people all week:

Start using your voice. Seriously. We interact with teammates via our voice, not our thumbs. If you're still typing everything, you're barely scratching the surface. The collaborative dynamic changes the second you start speaking to it. (I've written about this before—it matters more than most folks think. "Lose Your Thumbs, Find Your Voice")

Let Your Teammate Ask Questions. One of the biggest mistakes we make is assuming we've provided all the context the AI needs. We wouldn't do this with a human colleague—we'd expect them to ask for clarification if our request was unclear. We say, “My door’s always open if you get stuck.” Try adding this to your prompts: "Before providing an answer, please ask me any questions you need to better understand my request and deliver spectacular results." This simple addition transforms the interaction from a one-way command to a collaborative dialogue. The questions the AI asks will often reveal assumptions you didn't realize you were making, leading to significantly better outcomes.

Name your AI Assistants. Russ Summers, CMO at a tech start-up, has a budget GPT named Betty Budget. An advertising GPT called Aiden Adman. Revenue optimization? Roger RevOps. I know it sounds silly. But Russ told us something interesting on Beyond the Prompt: naming them changes how he shows up. "When I'm interacting with 'Betty' versus 'Budget Analysis Tool,' I naturally provide more context, explain my needs more clearly, and engage in more of a dialogue," he explained. "The results speak for themselves." Naming personifies AI just enough. Adds a little fun. Russ behaves differently because of that one stupid hack, and I bet you will too. Try it.

Block time to review its work. This one's counterintuitive. AI does in 20 minutes what used to take your team two weeks. Great, right? Except here's the problem: you're not ready to review two weeks of work in 20 minutes. Your ability to metabolize the information is now the bottleneck. So whenever you ask AI to do something meaningful—a market research study, a strategic analysis, whatever—block time on your calendar to actually review it. Just like you would if a human was doing the work. Otherwise the conversation gets buried and that creative spark you were courting? Snuffed out.

Actually onboard it. Quick diagnostic: If you had a new teammate who could write your all-staff memos, how would you onboard them? You wouldn't just say "write me a memo" and expect perfection, right? You'd show them examples. Give them your company's style guide. Provide feedback on drafts. Have you onboarded AI like that? If you had a teammate who could analyze your reports, you'd explain what insights matter, what your stakeholders care about, how you like information presented. Same with AI. It improves with guidance. It's a teammate. Treat it like one.

Here's Where I'll Leave You

Every day you treat AI like a tool instead of a teammate, you're leaving massive capability on the table. Not because AI can't do more—but because you're not setting it up to succeed.

What's one conversation you'd have differently today if you believed your AI was as capable as your smartest colleague?

Because I'm telling you: it probably is.

You're just treating it like it's terrible. Like it’s a tool.

Related: Lose Your Thumbs, Find Your Voice
Related: Beyond the Prompt: How to 3x Your Output with a GPTeam with Russ Summers

Join over 27,147 creators & leaders who read Methods of the Masters each week