Mistral has expanded Studio, its platform for building enterprise AI systems, with managed agents and apps as well as new workflow, context, observability, and evaluation capabilities. Teams can connect models such as Mistral Large 4 to company data and tools, then deploy in Mistral’s cloud, a private VPC, on premises, or across mixed environments.
Managed agents can carry out multi-step tasks using reusable skills, code sandboxes, subagents, and hooks for custom execution logic. Developers can set boundaries around the tools, data, and actions available to an agent. Mistral gives incident investigation as one example: an agent can follow a runbook, request approval before sensitive actions, and post findings in Slack.
Studio workflows combine fixed rules, model reasoning, and human decisions. Interrupted runs can resume, while event triggers and a visual graph expose how each process operates. Apps receive hosting, authentication, and access controls, and can use agents, workflows, search indexes, and shared enterprise context. Developers can test locally and deploy to Studio with one command.
For governance, OpenTelemetry traces record model calls, tool use, and approvals. Evaluations let teams test changes to prompts, models, or tools against their own datasets before release. Mistral says organizations including ABANCA, BNP Paribas, CMA CGM, and Tata Consultancy Services already use Studio in production.