Studio overview

Viete Studio / Agents

Your agents. Your rules.

Define purpose, shape behaviour, and connect the context your agents need. Then observe, evaluate, and refine. Using evidence from real work.

2.1.1 / Agent configuration

Define purpose, shape instructions, and choose the tools and sub-agents available to each agent. Configuration is how you control what an agent does and how it works.

2.1.2 / Data & resources

Connect the business context, knowledge, and resources your agents need. Execution quality depends on what the agent can access, not just how it is instructed.

Bring your data into the work

Connect relevant business data, documents, and reference material. Keep authorization and access boundaries part of every connection.

Reusable knowledge

Organise styles, skills, and domain knowledge as shared resources. A consistent starting point instead of rebuilt context per task.

Match models to the work

Choose from the models enabled in your deployment and assign them to agents. Match reasoning needs, cost, and response time to the task.

2.1.3 / Observe

Follow recorded agent activity: tool calls, conversation history, and sub-agent activity. Understand how a run unfolded before deciding what to change.

2.1.4 / Evaluate

Review costs, success rates, and operational patterns across runs. Use the evidence to prioritise what your team should investigate and refine next.

Operational performance

Costs, token usage, and success rates across available run data. These measures describe execution, not the correctness of the result.

Breakdowns by model and agent

Review performance by model, agent, or task. Find the work that deserves a closer look.

Improve deliberately

Make a focused change, run again, and inspect the result against your requirements. Refinement is a cycle.