The July end release introduces new capabilities that make agent interactions more intuitive, knowledge management more efficient, and external integrations more resilient. From richer chat experiences with familiar form widgets to centralized management of knowledge source connections and protection for external systems through rate limiting for MCP tools, these updates help teams build, manage, and scale AI solutions more effectively.
Interacting with agents is now more user-friendly with support for the same input widgets you already use in forms. In this preview release, widgets can be triggered through Python tools, allowing agents to collect and display information using structured, purpose-built interfaces directly within chat conversations.
Available for public preview, this enhancement creates a more guided and consistent user experience, helping users provide accurate information while reducing friction during agent interactions.
See more about Response types and UI widgets.
Manage knowledge source connections from one place
Availability: AWS and IBM Cloud only
Managing knowledge source connections is now simpler with centralized connection management. You can create, edit, and reuse knowledge source connections directly from the Connections Manager, providing a single location to manage connections across your workspace.
Previously, connections were configured separately within each knowledge source. With this update, you can:
This streamlined approach reduces duplicate configuration efforts, improves visibility into connection usage, and makes ongoing maintenance easier for workspace administrators.
Protect external systems with a rate limiter for MCP tools
To help protect external systems from excessive traffic, you can now configure a Rate Limiter control for MCP tools. This control restricts how many requests agents can send per minute and helps prevent service overload, API throttling, and unexpected spikes in tool usage.
When a configured limit is reached, the MCP tool call is immediately blocked and an error is returned. You can apply limits at both the tool level and the tenant level, and each limit is evaluated independently for every request.
This added layer of governance helps ensure reliable integrations with external systems while giving administrators greater control over agent-driven tool usage.
Looking for more updates?
These updates are some of several new capabilities introduced in this release. For a complete overview of the latest features, improvements, and fixes, see the What's New documentation. Staying up to date with the latest release notes helps you take full advantage of new functionality and platform enhancements.
Not using IBM watsonx Orchestrate yet? Start your free trial now.