Mistral expands Studio with managed agents, workflows, and enterprise context features

Mistral Studio now supports managed agents, workflows, and apps, allowing teams to deploy AI solutions across cloud, VPC, or on-premises environments while retaining full control over data and infrastructure.
The platform integrates with Mistral Large 4 (Le Chonk) to handle advanced coding, cybersecurity, and document reasoning, with new features for enterprise context, agent observability, and evaluations.
Internal use cases demonstrate significant efficiency gains, including an incident investigation agent that saves engineers hundreds of hours per month and a workflow that saves product operations teams approximately 35 hours weekly.
Customers such as ABANCA, BNP Paribas, and Capgemini are using the platform to automate critical processes like KYC onboarding and cargo release, with new processes becoming production-ready in weeks rather than months.

Platform Capabilities

Managed agents can be equipped with skills, sandboxes, and subagents to perform multi-step tasks, with hooks allowing custom logic at key execution points.
Workflows coordinate execution and approvals, supporting event triggers from enterprise systems and a visual graph for inspecting logic.
Apps provide hosting, authentication, and access controls, enabling developers to deploy applications with pre-approved components to reduce security review time.

Deployments can be hosted in the customer's own cloud, VPC, or on-premises, or combined with Mistral-hosted environments to meet specific security requirements.
OpenTelemetry traces provide visibility into model calls, tool use, and approvals, while evaluations allow testing changes against custom datasets before release.