Process Street is transitioning from a traditional interface-driven compliance platform to an agentic infrastructure by removing the necessity for manual user interaction. By releasing 105 new public API endpoints and mirroring them as tools on a hosted Model Context Protocol (MCP) server, the company is enabling AI agents to build, manage, and execute complex workflows directly. This strategic pivot targets enterprise environments where AI agents, rather than human operators, drive operational logic. The move positions Process Street as a programmable compliance layer that integrates into existing AI ecosystems like Claude and ChatGPT without requiring users to leave their primary work surfaces.
105 New API Endpoints and MCP Server Deployment
The company has expanded its technical capabilities by deploying 105 new public API endpoints, specifically designed to facilitate end-to-end workflow authoring. These endpoints allow for the creation, updating, duplication, and publishing of workflows, alongside the management of logic rules, assignment rules, due date rules, and widgets. Beyond basic execution, the update covers Pages, Data Sets with incoming webhooks, and a paginated workflow run listing that eliminates a previous 200-result cap. Crucially, the company is wrapping these endpoints in a Model Context Protocol server at mcp.process.st, effectively turning every API endpoint into an agentic tool. This allows an assistant to not only run a process but to actively build one, such as creating a compliance review or loading vendor lists into a data sets.
Agentic Integration with Claude and Microsoft Copilot
Process Street is positioning its new headless architecture to work across diverse AI environments, including Claude, Microsoft Copilot Studio, and various ChatGPT tiers ranging from Pro to Enterprise. The platform provides one-click connectors for Claude, ChatGPT, and Copilot Studio, while other clients can utilize bearer token authentication via a Process Street API key. A critical component of this rollout is the preservation of existing security boundaries; the company states that the API and MCP server inherit the current permission model. Admins, Builders, and Users can only interact with resources their existing credentials allow, ensuring that agents operate within established organizational constraints. However, the company noted that organizations utilizing SAML SSO are not currently supported.
Key Takeaways
- Process Street released 105 new public API endpoints to enable end-to-end workflow authoring and management by software and AI agents.
- The company launched a hosted Model Context Protocol (MCP) server at mcp.process.st to expose these endpoints as tools for AI assistants.
- The new infrastructure supports integration with Claude, ChatGPT (Pro, Team, Business, Enterprise, and Edu), and Microsoft Copilot Studio.
TechInsyte's Take
In our view, Process Street is making a decisive bet that the future of compliance operations lies in "agentic" rather than "manual" workflows. By moving to a headless model, they are acknowledging that the primary user of a compliance platform may soon be an AI agent rather than a human clicking through a dashboard. This shift is strategically significant because it embeds compliance directly into the developer and AI workflows that enterprises are currently scaling. By leveraging the Model Context Protocol, Process Street avoids the trap of vendor lock-in, allowing their logic to be called by any compatible agent. This is a sophisticated move to become the "operating system" for automated, auditable compliance.
Questions & Answers
How does the new headless model impact existing enterprise security and permissions?
The transition to a headless model does not introduce a new permission structure. Every one of the 105 new endpoints inherits the existing permission model, meaning AI agents can only access and manipulate resources that the specific user's credentials already permit.
Which AI platforms are compatible with the Process Street MCP server?
The MCP server is designed to work with Claude, Microsoft Copilot Studio, and ChatGPT (including Pro, Team, Business, Enterprise, and Edu plans), as well as any other application that supports the Model Context Protocol.
What specific workflow capabilities can an AI agent now perform via the API?
Agents can perform end-to-end tasks including creating, updating, and publishing workflows, managing tasks and logic rules, handling data sets via webhooks, and configuring permissions at the folder, workflow, page, and individual task levels.
Are there any current limitations regarding enterprise identity management?
Yes. While the platform maintains ISO 27001, SOC 2 Type II, HIPAA, and GDPR compliance, the company stated that organizations with SAML SSO enabled are not supported at this time.
Source: EINPresswire
