Here's the $30B mistake almost every company is about to repeat with AI:
They confused their product with their capability.
In 1990, C.K. Prahalad and Gary Hamel made the distinction famous in Harvard Business Review: a company isn't a portfolio of products. It's a portfolio of capabilities.
A real core competence passes three tests:
- It opens access to many different markets.
- It adds real value to the customer.
- It's hard for competitors to copy.
Watch how it plays out:
Honda's competence was never in cars. It was small engines, which is why one company sells motorcycles, lawn mowers, generators, outboard motors, and cars off the same skill set.
Canon was never a copier. It was optics and microprocessor control, which became cameras, scanners, and laser printers.
Kodak was never film. It was chemistry and thin-film coating. Eastman Chemical, the spin-off built on that capability, still thrives today. Kodak, the company that clung to the product, went bankrupt.
Same lesson, three times: the product is what you sell. The capability is why you win.
And it matters more in 2026, not less.
McKinsey's AI Transformation Manifesto (April 2026) is blunt about it: the advantage of AI leaders doesn't come from the tech, because those tools are available to everyone. It comes from how, and how fast, they apply it. Its first theme is "enduring capabilities over one-off solutions."
AI is a tool any competitor can buy. The capability to use it better than they can is not.
So before you green-light another AI pilot, answer Prahalad's real question:
Not "what do we sell?" but "which capability, if a competitor copied it tomorrow, would cost us everything?"
Most leaders name a product. The sharp ones name a skill.
Which is yours? And is your AI strategy compounding it, or just renting tools that make you look busy?
