Why we don't have a model picker

A question we get from customers (and occasionally from investors) is why we don't expose a model picker the way Cursor does.

The short answer is that we're building for a different user. Cursor's users are developers; picking between GPT-5 and Claude is a meaningful choice they're equipped to make. Our users include CFOs, sales leaders, customer success reps. For them, model choice is friction at best, and at worst a new vector for self-doubt. A marketing director second-guessing a revenue number because they're not sure whether they picked the right model is exactly the failure mode we're trying to design out of analytics.

A customer recently told us, half-joking, "I still have PTSD from Opus 4.7." It captured the dynamic perfectly. Even AI-native users don't actually want to be the one accountable for picking the model behind every answer... they want the answer to be right.

So we abstract it. Under the hood, we route to the best model per task: chart generation, chat response, and SQL writing each have different optimal models, continuously benchmark against our own evals, and swap silently the moment something better ships.

The user never knows. The user doesn't have to know.

This decision tells you where we think the moat is. The models are the commodity layer of this business. They'll keep getting smarter, cheaper, and more interchangeable. What doesn't get commoditized is the governance layer above them: the semantic definitions, the row-level permissions, the context architecture that turns a frontier model into a number a CFO will put in front of the board.

When the floor rises across model providers, we benefit. Our customers get smarter answers automatically, with no procurement cycle and no migration.

That's the right side of model commoditization to be on.