Here's the uncomfortable truth: AI can already build financial models better than most analysts. It can pull data, structure it, run scenarios, generate reports. It does this faster and with fewer errors. If you're a financial analyst whose job is "model building," you should be deeply nervous.
But here's the thing nobody talks about: The hard part isn't the modeling. The hard part is knowing what to model. What variables matter? What assumptions should you test? What's the story you're trying to tell? What matters for the decision that actually has to be made? That's not a technical problem. It's a judgment problem.
Senior analysts have always had an advantage—they know which questions matter. They know what analysts get wrong 80% of the time. They've seen how forecasts break against reality. They have pattern recognition about what to worry about and what to ignore. That's becoming their only real advantage, because the technical work—the modeling, the data pulling, the scenario running—is automatable.
The vulnerability is structural: AI + junior analysts might actually be a better combination than senior analysts alone. Cheaper, faster, fewer bad habits. The only thing that saves senior analysts is if they own the judgment layer. If they're deciding which analyses matter, which assumptions to test, what actually needs to be modeled for this decision.
The job market is reorganizing around this. You're seeing firms experimenting with structures like "experienced partner + AI + juniors" instead of "experienced partner + senior analysts + juniors." The experienced partner drives judgment. The AI handles execution. The juniors do research and ground-truth checking. The senior analyst role—the one between experienced partners and juniors—gets squeezed out.
Your concrete advantage: If you're a senior analyst, your job is becoming clearer—own the judgment layer explicitly. What questions does this business need answered? Which ones have measurable implications? How would we know if our analysis was wrong? What would change our mind? These are the questions that add value. The modeling is scaffolding.
If you're a junior analyst, the path is accelerated. You can do higher-level work faster with AI handling the mechanical part. Your advantage is learning to think like a senior analyst without spending 15 years as a modeling robot. Learn to ask good questions. Learn to understand what a decision actually needs to know. Learn to spot where analyses fall apart against reality.
The concrete action: Stop optimizing for modeling speed. Optimize for judgment speed. What's the minimum analysis needed for this decision? What would change the conclusion? What are we probably wrong about? These are your leverage points.
"In a world where anyone can model anything instantly, the scarce skill is knowing what's worth modeling in the first place."