Why AI Won't Replace (Most of) You

There's a lot of fear about AI right now — specifically, the fear of being replaced. I'm here to tell most of the professionals in my network: it's not going to happen. The value of experience and subject matter expertise is about to be more powerful than ever.

I've spent the past month or so obsessed with Agentic Engineering. My Anthropic bill is now considered an essential utility — prioritized higher than our car payments (in my mind, definitely not hers). In that time, I've built some genuinely impressive things.

And it has taken every single ounce of my experience and decades of accumulated tradecraft.

Here's the point: if you want to build something great with AI, you need someone with deep expertise in what you're building.

I work with a brilliant data analyst. We recently had a design session around some AI-powered analytics I'm developing in our space. Her presence — her expertise — was essential. Let me tell you why.

On the surface, I'm trying to ingest about a year of landing page performance data into my system so we can look for opportunities. Initially, we'll replicate the kind of trending, segmentation, and engagement reporting we use today. In the near future, we're going to do some next-level things. But we have to start somewhere.

That starting point was me loading a year of data from a bunch of CSV files. We asked the LLM for a trended view of visits and sales engagement on a specific landing page, and got a nice graph that told a story. During the past year, there were several clear inflection points — some spikes, some dips. A lot of it made sense.

But a few things surprised us. She and I hadn't looked at this page's performance in a while. We saw patterns that, had the page been getting more traffic, might have indicated problems or opportunities. But what caused those changes? We suspected shifts in paid media or CRM traffic. Sure, we could go look it up — but even then, why did that traffic change? Strategic pivot? Change in funding? Traffic quality shift? Competitive response?

That is where the LLM is going to help us in the near future — epic correlation and increasing causality — giving us a punchlist of exactly what the customer is looking for, how we can best engage them, and what drives good business outcomes. "What to do" will reveal itself. Roadmap items will be an embarrassment of riches. Implementation will be guidance, not labor.

That destination IS coming. But knowing how to work with data to extract insights is going to be fundamental to getting there. And there is no way we're getting there without our analyst. Here's why.

We got our trended graph. We saw the peaks and valleys. But we want to know why. She either knows the answer or knows exactly how to find it. For instance, she would:

• Look for correlation with promotional activity (which requires knowledge of our promotional constructs, cycles, and economics)
• Check for changes to the page itself — content optimizations, creative refreshes, redesigns, updates to product offering
• Investigate any multivariate tests that might have been running
• Connect with demand generation teams (Media, CRM) on potential strategy or budget shifts
• Reconcile known data quality issues, gaps, or historical events like re-attributions or metric redefinitions
• Factor in industry or product-specific seasonal trends
• Cross-reference segments to look for patterns of similarity or difference
• And probably 20 other things I don't know, because I'm not a data analyst

My AI system — all AI systems — need that training. They need that context, those lenses, and an understanding of what good looks like. And it will always need to evolve: new technology, new risks, emergent opportunities, shifts in the marketplace. But here's the thing about all that glorious plate-spinning — most of it becomes orchestration, articulation, and evaluation. Which means we'll have bandwidth. Bandwidth to move faster. Bandwidth to increase scope. Maybe even bandwidth to invest more in customers, employees, and the world beyond pure profit motives — because we'll have the surplus.

Those with deep experience, strong hard and soft skills — they're going to thrive. Good managers and communicators are going to thrive. Good organizers and data-driven thinkers are going to thrive. And it won't even matter if you catch the first wave. It will catch you. My dad sends text messages now. For the longest time, I thought that would never happen. Eventually I realized it was inevitable.

So worry not. If you know your stuff, you're going to be just fine. There's a tricky question down the road about how the next generation gains the knowledge we built through manual labor — but we've got some time. I'm pretty sure we'll figure it out.

We'll have help.