AI made B2B products easier to enter. It didn't make customers free to serve.

In a November 6, 2026 post^{[1]November 6, 2026 post}, Jason Lemkin argued that B2B companies still have defensive advantages, but those advantages are much less protective when strong competitors can emerge quickly with AI. It is a warning incumbents should take seriously.

The mistake is to confuse lower entry costs with lower costs of serving customers. Those costs can move in different directions.

Let F be the fixed cost to create a useful application. Let m be the marginal cost of supporting one customer each month. Across N customers, average cost is approximately F / N + m.

AI can reduce F, perhaps dramatically. That makes small and specialized markets viable where previously they weren't. But m still includes hosting, support, security updates, external API changes, customer training and the consequences of product failure.

A company with ten customers and an excellent fit might be competitive despite having fewer engineers. A large incumbent may still have valuable integrations, data contracts and service capacity. Neither automatically wins.

Aurekai's model is interesting because several products can share part of F and m. A new business may use established identity, CRM, finance, media and data systems. Its customer-facing experience can remain distinct without rebuilding everything underneath.

That is an economy of scope: the same capabilities contribute to multiple products. The advantage isn't a moat made of code that competitors cannot copy. It's an operating organization that can keep developing and servicing products economically.

Lemkin's prediction of faster competition seems plausible. The better response is to improve the parts customers actually rely on and make it easier to change systems when necessary. Software companies will need better service, not just more difficult-to-replicate screens.