From “Using” to “Teammate”

How Professor Jeremy Utley Changed My AI Game

Until recently, I thought of AI the same way most people do: as a tool—something you use to get from Point A to Point B more efficiently. Whether it was drafting emails, summarizing documents, or generating ideas, my mindset was focused on inputs and outputs. Treat it like a smarter search engine or a faster assistant, and move on. Simple.

But then I attended a workshop led by Professor Jeremy Utley, and something shifted. Not immediately, and not because of some grand revelation—but through a series of thoughtful, well-crafted exercises that challenged how I approached creativity, problem-solving, and collaboration in the age of AI. What struck me wasn’t just the practical techniques he shared, though those were valuable. It was the subtle but powerful idea at the core of his teaching: if we want AI to elevate our thinking, we have to stop treating it like a tool and start treating it like a teammate.

That idea landed harder than I expected. After all, in any decent workplace, we don’t talk about “using” our colleagues. We collaborate with them. We challenge each other, refine each other’s thinking, and build better outcomes together. Why should it be different with AI—especially as it becomes increasingly capable of reasoning, questioning, and even challenging our assumptions?

What followed for me was a mindset shift: away from control and toward cooperation. From commanding an interface to engaging with a partner. And that shift has changed not just how I interact with AI, but also the quality of the work we produce—together.

And don’t get me wrong, I am still one of these AI enthusiasts adding “please” and “thank you” in my prompts and interactions but it was not a mindset, just a rule of etiquette.

1. AI as a Co-Worker, Not a Calculator

Utley’s mantra “Teammate, not technology” cracked my mindset like a Monty Python sketch gone meta . In the same spirit, he reminded us—it’s absurd to say “I used AI” the way you’d boast, “I used a hammer.” A better question would be: Did I collaborate with every resource I have to make this the best work possible?

2. Make AI Ask Questions

One of Jeremy’s hallmark techniques: before AI answers, it should ask you clarifying questions . Just like a colleague checking you’re aligned, AI probing for context avoids hallucinations and surfaces hidden assumptions. That workshop moment when GPT paused and asked me, “Can you clarify your goal?”—that felt like a teammate stepping in, not a black box spitting code. Or as Professor Utley put it, AI should be an inquisitive interviewer trying to go to the deeper level of your thought process.

3. Chain of Thought = Shared Reasoning

Hey, humans reason out loud—and Jeremy insisted AI can, too, if prompted! Asking it to walk through its logic produced dramatically better results in his dad’s PDF-to-CSV example . I tried it: “Before answering, show me your thinking step-by-step.” Boom—my AI partner went from random text soup to a crisp, structured CSV. Like pairing with a junior colleague who explains their math instead of dropping spreadsheets.

4. Critique, Feedback, Iterate

Utley emphasized: never take AI’s first draft as gospel. Treat its output like you would an intern’s work—review, comment, refine, and ask for the next draft . That subtle shift—“Good start; could you try five more like this, more punny”—turned outputs into exactly the tone and quality I wanted.

5. Voice, Multi-Agent Workflow

One workshop highlight: Jeremy spoke about his writing setup—he talks to one AI, types responses, routes its answer to another for critique, and even uses voice-mode to brainstorm verbally . He joked it’s like dictating to oneself in the bathtub like Churchill. Beautiful chaos—and brilliant collaboration.

🚀 Concrete Wins in My Day Job

• Engineering Documents: Now, instead of dumping a stack of specs into AI with “Summarize,” I tell it: “You’re a product manager. What clarifying questions do you need before summarizing?” The result? Cleaner, context-aware briefs—with fewer follow-up edits.

• Design Brainstorms: No command-line requests, but real back-and-forth: “What if we tried X instead of Y?” Then: “Great—expand three top options with risks & benefits.” It’s like pairing with a junior designer.

• Email War Rooms: Instead of “Write polite email,” I say “Write as my empathetic peer. But here’s how I don’t want it to sound.” AI asks, I refine, I approve.

🧠 The Takeaway (AKA “Don’t Be a Tool-User”)

Professor Utley isn’t selling snake oil—he’s selling a mindset:

1. Treat AI like a team member, not a shortcut.

2. Ask it to think aloud, not just spit answers.

3. Encourage dialogue, not just monologue.

4. Critique and iterate like any human teammate.

5. Use voice and multi-agent workflows to mimic real-life collaboration.

The result? Outputs that feel more insightful, original, and on-brand—built with intention, not just generated. Stop “using AI” like a hammer. Instead, pull up a chair and invite it in over coffee. (Or, you know, shampoo and shower thoughts.)