Workbench now helps teams design and configure their data exchanges, not only operate them at runtime. Provisioning Onboarding has been available in Beta since the end of August. Administrators build form templates from SecureTransport configuration entities, then invite partners and business users by email to fill in the fields assigned to them. Workbench deploys the resulting configuration directly to SecureTransport. Partner onboarding becomes a guided, collaborative workflow where every step is tracked. Contributors don't need SecureTransport expertise, and configuration details no longer travel by email. The Beta capabilities are off by default and can be turned on when you are ready to try it. 

Following its Beta introduction last quarter, the Workbench MCP (Model Context Protocol) Server is now generally available. Customers can connect the AI assistant you already use to Workbench and work with your file transfer operations through natural-language questions and supported actions. Every request runs with the permissions of the connected user or service account. The assistant only sees and does what that identity is already allowed to, so AI adoption builds on the access controls you already trust.

Three Beta capabilities let customers go further today:

• Bring your own LLM (Beta): customers use the model provider their policies already approve, such as AWS Bedrock, Azure Foundry, Anthropic or an OpenAI-compatible model. This removes a common blocker in regulated accounts that won't accept a provider imposed on them.

• Embedded MCP Client (Beta): puts AI directly in the Workbench interface, so operators can ask questions and investigate issues where they already work, with less reliance on a few MFT experts.

• MCP Routing for fleet management (Beta): lets one AI connection to Workbench discover and use the tools exposed by the SecureTransport and Transfer CFT instances registered in Workbench, so customers no longer need to connect to each product separately. This makes Workbench the single AI entry point for managing a whole fleet of transfer products, and a strong opener for AI-readiness and modernization conversations with customers running many SecureTransport and Transfer CFT instances.

Customers can now create their own dashboards from predefined data sources, extending Workbench's built-in views. Each team can focus on the indicators that matter to its role, such as operations tracking transfer activity and errors, or business teams following the transfers they're responsible for. Everyone gets a clearer, faster view of what needs attention, and can drill down into the details to investigate further.

Workbench now fits more easily into customers existing identity setup. Alongside SAML, configure OpenID Connect identity providers, so users sign in to Workbench through the corporate single sign-on they already use. SAML setup is also simpler with providers such as Okta. Administrators can define the NameID format sent in signed requests, so Workbench aligns with the identity provider's expected configuration without workarounds. Access to Workbench then follows corporate authentication policies, including multi-factor authentication enforced by the identity provider, and there's no separate set of credentials to manage. The internal identity provider remains available as a fallback for administration and diagnostics.

Administrators now have a single Data Purge page to see and manage how much space Workbench data takes up. It shows total storage and the next scheduled purge at a glance. For each tracked data table, such as transfer tracking, logs, audit and routing information, administrators see its size and row count, with the largest table flagged. For each table, they can set a retention period, from days to years, and a purge schedule. Automatic purge can be switched on or off per table, and a purge can also be run on demand. The latest purge status confirms that each purge completed. Administrators keep the history they need, such as long audit retention for compliance, and clean up high-volume tracking data sooner. This keeps database growth predictable without manual cleanup or database scripts.