At Inspo we design and build software products — and the intelligence inside them — shipped end to end by one team.
A wrapper takes a frontier model, adds a prompt and a interface, and sells the gap between what the model can do and what its buyers know it can do. That gap is closing fast. What remains when it closes is the work wrappers skipped.
The line, precisely
It is not about training models from scratch — almost nobody should. The line is ownership of the system around the model: the data pipelines that feed it, the evaluation that constrains it, the actions it can take, and the product it lives inside. Wrappers own a prompt. Builders own a system.
Why the difference compounds
A wrapper improves when the model provider ships. A system improves every week you run it.
A wrapper’s knowledge is generic. A system accumulates your data, your edge cases, your tone.
A wrapper can be cloned in a weekend. A system is a moat made of operations.
What we do instead
We build the system: agents wired to your stack, automations that act in your CRM and your inbox, evaluation that catches regressions before your customers do. The model inside is replaceable — that is the point. Everything around it is yours, and that is the product.
Let’s keep in touch.
More notes from the studio on X and Instagram — product, design and AI, from the people shipping it.
