Your MBA Won't Save You. Your Experiments Might.

I watched my client's face when I told him I could build the system in Claude instead of the custom GPT he'd been using for six months.

He didn't ask me what Claude was. He didn't ask me how it worked. He asked me one question: "Will it be better?" And honestly, I wasn't sure yet. I just knew his current setup was producing mediocre restaurant email campaigns that all sounded the same, and I'd been experimenting with something that didn't. So I built it. It was better. And that single moment changed my entire trajectory at that company.

TLDR

• The consultants winning right now aren't the ones with the best credentials. They're the ones who've tried everything and know what actually works.
• McKinsey says 57% of U.S. work hours are technically automatable with tools that exist today. The people who understand those tools have the advantage.
• Experimentation velocity beats expertise depth when the landscape shifts every few months.
• Your 10 years of experience matters less than what you shipped last Tuesday. Is that fair? Probably not. But it's real.
• The gap between "I know about AI" and "I've built something with AI that's been running for months" is where all the money lives.

Why AI Skills for Consultants Now Mean Something Different

There's a running joke in AI circles that Claude launches an update every week and half the jobs in the industry vanish overnight. It's not entirely a joke.

McKinsey's November 2025 report found that current technology could automate roughly 57% of U.S. work hours. Not by 2030. Not if AI keeps advancing. With what exists right now. For context, their 2023 estimate was 30% by 2030. The number nearly doubled in two years.

So here's the question every consultant, coach, and knowledge worker should be sitting with: if the tools change this fast, what exactly is your expertise built on?

Because if it's built on knowing a specific platform, a specific process, or a specific way of doing things, you're standing on something that shifts every quarter. The consultants I see thriving aren't the ones with the deepest knowledge in one thing. They're the ones who've touched fifteen things, broken twelve of them, and figured out which three actually work for a given client's situation.

That's not a credential. That's an experiment log.

The Experiment That Changed My Job Title

I was hired as a senior AI copywriter. That was the official job description. The client ran a top restaurant coaching business, with group coaching, one-on-ones, masterclasses, the whole setup. He had years of content sitting in recordings, transcripts, frameworks. Mountains of it. Mostly untouched.

The previous approach was a set of custom GPTs that produced passable outputs. I could have kept running those. Nobody would have complained. The outputs were fine. Fine is the enemy of good when your client's brand depends on sounding like a specific human being.

So instead of sticking to the script, I started experimenting. I built AI skill systems, not prompts. Systems with verification loops, reference libraries, voice encoding, quality gates. Things that could process a coaching transcript and produce a full email campaign that sounded like the client, not like a chatbot wearing his jacket.

Within four months, my title changed from senior AI copywriter to head of marketing and AI. Not because I asked for a promotion. Because the scope of what I could deliver kept expanding as I kept experimenting. The World Economic Forum projects that 92 million jobs will be displaced by 2030 while 170 million new ones emerge. The new roles go to the people building, not the people waiting.

Why Credentials Are Losing to Proof of Work

Here's what I think is actually happening. And I mean this genuinely, not as some provocative LinkedIn take.

The value of a credential used to be that it signaled depth. You spent years learning something, and the degree or certification proved it. The problem is that the thing you spent years learning might look fundamentally different in six months. Research from the Dallas Fed shows that AI is already substituting for entry-level workers who bring primarily textbook knowledge, while it complements experienced workers who bring judgment that can't be codified.

That's the split. Codifiable knowledge, the stuff you learn in courses and certifications, is exactly what AI replicates fastest. The tacit knowledge, the judgment you develop by actually doing the work across messy real-world situations, that's what still holds value.

But here's the uncomfortable part. You can't develop that tacit knowledge by studying. You develop it by experimenting. By building something, watching it fail, understanding why it failed, and building the next version. I've done this over 50 times now. Fifty-plus AI skills, each one teaching me something the previous one didn't cover. V1 of the brand blueprint skill missed positioning gaps I caught manually. By v14, the skill was catching gaps I was missing. That doesn't come from a curriculum. It comes from reps.

The Coordination Layer Nobody Talks About

There's another piece to this that goes beyond individual experimentation, and it's the part I think gets overlooked the most.

I don't just build skills. I coordinate between team members, manage deliverables, run a customer success function, and make judgment calls about which system to deploy for which problem. That coordination layer is where the actual value sits.

Think about it this way. The skill file does 80% of the production work. But deciding which skill to use, knowing when the output needs a human override, understanding what the client actually needs versus what they asked for — that's the coordination layer, the 20% that AI can't replicate well yet. McKinsey's same report found that over 70% of skills employers want today are used in both automatable and non-automatable work. The skills don't disappear. How they're applied changes.

I logged every hour of my work for a week once. The revelation wasn't how much time I spent building. It was how much time I spent making contextual judgment calls. Which tone works for this client's audience. Whether the system's output needs a full rewrite or a light edit. When to push back on a client's brief because what they asked for isn't what they need. None of that is in any job description. All of it is why I'm still useful.

So What Does This Actually Look Like for You?

If you're a coach, a creator, or any kind of knowledge worker reading this and wondering what to do with all of this, I think the answer is simpler than it sounds.

Stop trying to learn AI in the abstract. Pick one workflow in your business, the most annoying, time-consuming, repetitive one, and build something that handles it. Not a prompt. Not a template. Something with enough structure that it produces consistent results without you thinking about it every time. Demand for AI fluency in job postings has grown sevenfold in two years, and that demand isn't for people who took a course. It's for people who built something.

Yesterday it took me 30 minutes to write a 1,200-word blog and two LinkedIn posts. Published, SEO-optimized, cross-posted. It was 11pm and I'd been at it for six hours building the skill that made that possible. The 30 minutes is the output. The six hours is the experiment. The experiment is where the value lives.

The consultants of the future won't carry degrees. They'll carry proof of what they've built and broken and rebuilt. If you want to see what that looks like in practice, read about how I built AI production systems that serve 15 clients per cycle. And that's either terrifying or exciting depending on which side of it you're standing on.

If you're a coach or creator trying to figure out where AI actually fits in your business, not theoretically, but practically, that's the conversation I have with every client before we build anything.

Let's have it. Free 30-minute call, no commitment.