
TJ Walker argues that getting great AI output depends on communication skills, not technical tricks. Start every prompt with a one-sentence role description like "you are an experienced executive communications specialist." Use a three-layer follow-up to refine responses: state what to keep, what's wrong and why, and what should change instead.
Chapters
Today on Ultimize Discover, you're getting a short practical sample from the Ultimize universe. Use this on its own, take one idea and apply it today, or go deeper at ultimize.com. Ultimize is the world's largest personal development podcast, audiobook, and online learning network. Ultimize. Here's something most people never do, and it takes about 10 seconds. Before you write or dictate your next AI prompt, give it one sentence that tells AI what role to play, not what to write, who to be. Something like, you are an experienced executive communications specialist who writes for senior leaders in financial services. That one sentence before you've asked for anything fundamentally changes the world AI thinks it's operating in. It stops producing generic output and starts producing something calibrated to a specific professional context. Try it once. The difference is immediate and it's remarkable. That's the kind of thing we get into on Talk to the Machine, a limited series about the communication skills that make AI genuinely perform. Not the technical side, the The human side. Clarity, context, tone, iteration. The skills that have always separated great communicators from average ones now apply to the most powerful tool most of us use every day. Hi, I'm TJ Walker, and I'm your host. This is part of Ultimize, the world's largest personal development podcast, audiobook, and online learning network. I spent decades coaching executives, politicians, and business leaders on communication. The people getting extraordinary results from AI aren't the most technical people in the room. Typically, they're the best communicators. 10 sessions right here in this limited series podcast. Subscribe to Talk to the Machine wherever you get your podcast and share it with someone who needs an edge. Let's go. The best prompt I ever saw took 11 minutes. A client of mine, a senior communications director at a mid-sized financial services firm, typed this into an AI system. "Explain this to a nervous first-time investor in plain English." That's it, 11 words. And what came back was so clear, so useful, and so precisely calibrated to the right audience that she used it almost verbatim in a client memo the same afternoon. She didn't get lucky, she communicated well. Most people don't do this. Most people treat AI like a search engine. Type a few keywords, hit enter, and hope, hope something useful comes back. They ask it things like, "Write a summary," or "Help me with this email," or "Explain blockchain," with no context, no audience, no purpose, and no direction. And then they're disappointed by what they get. Well, of course they're disappointed. Vague in, vague out. That's not the AI failing, that's the communication failing. That's what nobody tells you when you first start using AI. The quality of what you get back is almost entirely determined by the quality of how you communicate going in. AI systems don't have intuition. They don't know what you mean when you don't say what you mean. They can't read between the lines, fill in the blanks, or infer your unstated goals. They definitely can't read your mind. What they can do, and do remarkably well, is respond to clear, specific, well-framed communication with equally clear and useful output. That's the entire game. I've been coaching executives, politicians, business leaders on communication skills for decades, and in the last few years I've watched something fascinating happen. The people who are getting extraordinary results from AI tools aren't necessarily the most technological people in the world. They're not the coders or the engineers or the prompt engineering specialists. They're the people who are good communicators. They know how to be specific, how to give context. They know how to frame a request so that anyone, human or machine, understands exactly what's needed. Think about what makes you a great communicator with another person. You don't walk up to a colleague and say, help me with the thing. You give them context. You tell them what the goal is, who the audience is, what tone is appropriate, what constraints you're working under, and what good looks like. Every skilled communicator does this instinctively. The only difference with AI is that you have to do it intentionally and in writing, or dictate it so it's turned into writing, because AI doesn't have the benefit of knowing you knowing your history or reading your facial expressions. All that's lost. The gap between a mediocre AI result and a brilliant one is almost always a communication gap. I want you to hold on to that idea because everything in this series is built on it. This isn't a program about technology, it's a program about communication, about the specific skills that have always separated great communicators from average ones. Now applied to the most powerful tool most of us will use in our professional lives. Let me give you one more example before we get into the mechanics. I had a client, a VP of marketing at a consumer goods product, who came to me frustrated with AI. He'd been using it for months and felt like it was consistently producing generic, bland content that didn't sound like him or his brand. When I asked him to show me how he was using it, the problem was immediately obvious. He was typing things like, write a social media post about our new product launch. And the AI was producing exactly that, a generic, bland social media post, because that's exactly what he asked for. He spent about 20 minutes reworking how he communicated with it. We added the brand voice. We specified the audience. We gave it the emotional tone he wanted to strike. We told it what to avoid. We told it the length, and we told it why this launch mattered to real people's lives. The next output, well, that was completely different. It was specific. It had personality. It felt like his brand. Nothing about the AI had changed, but his communication with the AI had changed. That's the fundamental insight this series is built on. AI is a communication partner, not a vending machine. You don't insert a coin and receive a fixed output. You have a conversation, or at least the opening of one, and the quality of that conversation is entirely in your hands. And yes, when I say conversation, I do recommend you literally speak to your AI. That's what I do almost all of my interactions with AI with. I'm speaking to it. It captures so much more emotion, nuance, and more detail. And I can speak faster than I can type. I don't know about you. Remember, the better you communicate, the better the AI performs. And the skills required to communicate well with AI are exactly the same skills that have always made some people more effective in every room they walk into. In the sessions ahead, we're gonna go deep on each of these skills: clarity, context, tone, iteration, constraints, perspective shifting, high-stakes preparation, and the honest limits of what AI can and can't do. By the time we're done, you'll have a system for communicating with AI that compounds over time, not just a few tricks that work once. But it all starts here, with the most important shift you can make. Stop treating AI like a search engine and start treating it like the most responsive communication partner you've ever had, because that's exactly what it can be if you give it what it needs. Most people use AI like a vending machine. They put something in, they get something out. They evaluate whether they like it, and if they don't, They either accept it anyway or they walk away disappointed. That's not a communication strategy, that's a transaction. And it's the single biggest reason so many people underestimate what AI can actually do for them. I was coaching a product manager at a midsize software company, a sharp, organized thinker who had been in product for about 8 years. He was using AI to help draft product requirement documents, stakeholder updates, and internal communications. His complaint was that AI kept missing the nuance he needed. It would get the structure right but flatten the complexity, or it would capture the detail but lose the narrative thread that made the document readable for non-technical stakeholders. He felt like he was spending as much time fighting the output as he would have he just written it himself. When I asked him to walk me through his process, I solved the problem in about 30 seconds. He was writing one prompt, getting one response, deciding it wasn't quite right, and then starting over with an entirely new prompt from scratch. He was treating every interaction as a single transaction rather than the opening move in a conversation. The iteration mindset starts with a simple reframe. Your first prompt is not a request for a finished product. It's an opening bid. It's you saying, here's roughly what I'm after. Let's start here and work towards something that's actually right. The best AI communicators I've observed think about interaction in rounds. Round 1 is about direction, getting AI oriented toward the right territory. Round 2 is about refinement, narrowing down what's working and what isn't. Round 3 is about precision, pushing it into the specific adjustments that take something from good to genuinely useful and occasionally great. That's 3 rounds, sometimes 2, occasionally 4, sometimes even more, but it's almost never 1. The key skill in iteration is being specific about what to keep and what to change. This is where most people fall down. When the first output isn't quite right, they'll say something like, "Make it better," or "Try again," or "Make it more professional." That's not useful feedback. That's the equivalent of telling a colleague their presentation wasn't good without telling them what specifically wasn't working. Effective iteration sounds like, "The structure in paragraph 2 and 3 is right. Keep that." "but the opening is too formal for this audience and the closing is burying the call to action. See the difference?" You could also say, "Rewrite the opening to sound warmer and move the call to action to the second sentence of the conclusion." Specific, directional, actionable. That's feedback AI can actually use. I want to give you a technique I call the 3-layer follow-up. I want you to use it on your very next AI interaction that doesn't hit the mark on the first try. After you get your first output, before you do anything else, answer 3 questions in your follow-up prompt: What's working and should stay? What's not working and why? And what would the ideal version do differently in specific terms? Write all 3 answers into a single follow-up message, or do what I do, which is just dictate into the AI, it's often easier and faster. Just 3 or 4 sentences total is fine, but include all 3 layers. What I've consistently seen is that this kind of structured follow-up produces a second output that's dramatically closer to what the person needed than a vague try again. The product manager I mentioned worked through this with me on a stakeholder update he was trying to produce. This was gonna be for a quarterly business review. His first prompt got him an output that was structurally fine, but too technical and too long. Instead of scrapping it and starting over, we wrote a follow-up that said, keep the 3-section structure and the summary table. Those are right. Shorten the whole thing by 30%. Rewrite the technical sections for a chief executive who understands business outcomes, but not engineering specifics. Translate every technical metric into a business outcome. And make the opening sentence something she'd actually want to read rather than a status update. The second output was, by his own assessment, 90% of what he needed. He spent about 5 minutes on final edits. What had been a frustrating 40-minute exercise became a 10-minute one. There's a deeper principle underneath the iteration mindset that's worth naming explicitly. Great communication has always been iterative. Nobody writes a great speech in one draft. Nobody produces a brilliant proposal without revision. Nobody gives a perfect negotiation argument on the first pass. The best communicators have always known that the first version is where you discover what you're actually trying to say. And the subsequent versions are where you say it well. AI doesn't change this reality. What it changes is the speed of each cycle and the ease of making substantive changes without starting from zero. The iteration mindset isn't a workaround for AI's limitations. It's the same creative process that's always produced great communication. Now running faster. One thing I want you to push back on is the instinct to feel like a failure when the first output isn't perfect. I work with very senior, accomplished people— chief executives, cabinet members, authors, executives at major institutions— and not one of them gets it right on the first prompt every time. The ones who get consistently great results aren't the ones with perfect first prompts. They're the ones who iterate with precision and confidence, who treat each round of output as information about what direction to move in next, and who understand that refinement is part of the process, not evidence that something went wrong. The product manager checked in with me a few months after our work together and said something that I thought captures the iteration mindset exactly. He said he stopped thinking about AI as a tool he used, and started thinking about it as a process he ran. The distinction mattered to him because processes have steps, and steps are things you can get better at. He was right. The iteration mindset isn't just a technique for better AI output. It's a shift in how you think about communication itself as something that moves, adjusts, refines, and arrives at something good through disciplined engagement. Rather than lucky guessing. By the way, quick question for you. Have you ever noticed how easy it is to consume content and how hard it is to actually do anything with it? I mean, you can listen to 100 episodes, read 10 books, and your habits are exactly the same. That's the gap most people never close. So if you're listening to this and thinking, yeah, I want to actually improve, go to ultimize.com. We've got free courses there for you, either on communication skills or building better habits, and it's designed to get you to actually what you're learning, not just hear it once. No cost, no catch, just go to ultimize.com and start. Ultimize.
More from TJ

The Breakout Cases - What Actually Caused Famous Speaking Career Breakthroughs | Ultamize

The Red Circle Sessions Finale - Build the Speaking Career You Came Here For | Ultamize

Two Stages, One Logo, Two Completely Different Realities - Main Stage vs Independent Events | Ult...

Pre-Reset - Why Repetition Builds Better Speaking Decisions | The Red Circle Sessions | Ultamize

Your Decision, Your Framework, and What Happens After - Build Your Speaking Career First | Ultamize

The Red Circle Sessions Trailer - The Speaking Career Truth Nobody on That Stage Will Tell You | ...

The Two-Minute Ultamize Reset - Your Platform, Your Audience, Your Voice | The Red Circle Sessions



