A tool teams rely on after every meeting should not itself depend on one thing that can break. If Sol routed every transcript to a single model from a single lab, then that lab's bad afternoon would become every customer's bad afternoon, on the day they most wanted their outcomes written up.
Not every step needs a frontier model
Turning a meeting into outcomes is not one call. It is a few steps, and they do not all need the same horsepower. Reading a long transcript for decisions and implied owners is the hard step, and it earns a frontier model. Writing a short summary, or matching this week's progress against last week's open items, is lighter work a fast model does well. Routing each step to the right model is how you get quality where it matters and a sane bill everywhere else.
- First-pass extraction routes to a strong model that reads the whole conversation.
- Summaries and cross-meeting matching route to a fast, inexpensive model.
- If a provider is degraded, the router falls to the next candidate that clears the bar.
Open models, on your own ground
Model-agnostic also means a company is not forced onto a hosted API. Sol can route to open models a team hosts itself, so transcripts never leave their perimeter. The routing table is the same. Only the endpoints change.
The router is a reliability feature. Labs move fast, and your meetings should not notice.
Frontier models will keep leapfrogging each other, and open models will keep closing the gap. A registry and a router mean Sol can take the best of each as it arrives, without a rewrite, and without asking anyone to care which lab answered.