Explore put the right intelligence behind the responsibility.
Every Monday, your customer notes need to become a concise account of what people are asking for. Once a quarter, those accounts help you decide which problem deserves the next product effort. The same workforce can support both responsibilities. You may want a different model behind routine preparation than behind the synthesis that informs a larger decision.
Supanova's model controls let you make that choice at the workforce level. Set the supported models available to your workspace, assign defaults to workforce bands, and use finer tier settings when a broad assignment does not fit. The people directing the work still decide what the output must prove and what action to take from it.
Picture an illustrative customer-insight project. A specialist organizes notes from recent support conversations into recurring questions, with the source behind each theme. Another specialist compares those themes with product research and prepares a decision memo. The first job rewards consistent extraction and clear labeling. The second needs room to weigh conflicting evidence and explain a tradeoff.
Memo under review: Customers keep asking where to find the status of a request. Support notes suggest the guidance is unclear; interviews also point to a missing view. Which explanation should the team test before it commits to building anything?
Your team might choose one configured model for focused preparation and another for coordination and synthesis. Each part of the work has a deliberate starting point, and your people can spend their attention on whether the memo answers the real question.
The visible finish matters. If the weekly account misses a repeated issue, correct the brief or the model assignment. If the quarterly memo keeps turning weak evidence into confident advice, inspect the sources, feedback, and revision pattern. Model controls are useful when they help you improve work you can actually judge.
The first decision is the Multi-model palette: which supported providers and models your workspace permits. Tier-band settings then offer eligible choices from that set. Enabling a model in the palette leaves existing assignments in place until you change them.
You can assign models by three workforce bands. Contributors often handle focused production. Leaders coordinate and synthesize. Executives handle broader strategic and cross-functional work. These are Supanova workforce tiers, not the titles of the people on your team. The assignment gives each band a configured model to begin with.
When one band covers work with different needs, finer tier settings let you configure allowed models, a default model, and a fallback model for a particular tier. They also expose request and daily model-budget limits. These choices follow the supported options and execution path available to your workspace.
This structure means you can make a considered change without redesigning every assignment. A team can give the recurring summary a stable starting configuration, then give a higher tier a different default for the quarterly synthesis. The exact supported choices and current cost indicators are shown in the workspace settings.
Model controls answer which model is configured for a part of the workforce. Nearby features answer other questions:
Choose a repeatable responsibility with a clear standard. For the customer summary, that might mean every important theme has supporting examples and unresolved contradictions are visible. Keep the source material and review standard comparable while you change one model assignment. Then look at first-pass acceptance, revisions, and the cost of a result your team can use.
The useful configuration fits the capabilities your work needs and produces acceptable results within the limits you choose. A deliberate model choice becomes part of how you give the workforce direction, alongside the brief and the people who review it.
Can I keep a fixed model assignment?
Yes. Tier-band and tier-level settings let you choose configured models. Dynamic Model Selection is a separate choice for supported work; without it, the configured assignments remain the starting point.
When should I use finer tier settings?
Use them when a broad workforce band covers responsibilities with materially different capability or cost needs. You can define a tier's allowed, default, and fallback models within the supported choices available to your workspace.
What should I check before removing a model?
Look for tier or tier-band assignments that rely on it. Move those assignments to another enabled, suitable model before removing the old choice from the workspace palette.
Start with one responsibility and the result your people would recognize as useful. Choose the supported configuration that fits the work, inspect what comes back, and refine from evidence rather than a model name alone.
Launch a new workspace
Previous: Multiple workspaces · Next: Agent progression · Back to the Feature Catalog
Launch a new workspace