The Process Street MCP server lets AI assistants operate directly on your workflows, runs, tasks, data sets, and permissions. Connect it to Claude, ChatGPT, or Copilot Studio and start managing operations through conversation instead of clicks. Here are 10 practical ways teams are using it right now.

The Model Context Protocol is an open standard that lets AI assistants connect to external tools. Process Street ships an MCP server that wraps our full public API, giving AI agents the same capabilities your team has in the UI: creating workflows, running checklists, updating records, managing permissions, and pulling reports.

No plugins. No custom code. Connect once, and your AI assistant can read and write to Process Street as if it were a team member with the right permissions.

1. Build workflows by describing what you need

Skip the blank canvas. Tell your AI assistant what the process looks like in plain English, and it builds the workflow for you: tasks, form fields, conditional logic, approval gates, assignment rules, due dates.

Example prompt: “Create a vendor onboarding workflow with 4 steps: collect W-9 and tax ID, run a compliance review with a risk tier dropdown, add a manager approval gate, then activate the vendor in our system. If the risk tier is High, show an extra compliance notes field.”

The MCP creates the workflow as a draft, lets you review it, and publishes only when you confirm. Full CRUD on workflows, revisions, tasks, and widgets shipped with our 105-endpoint API release.

2. Get a morning operations briefing without opening a dashboard

Ask your AI assistant for a status update across your workspace. It reads your active runs, overdue tasks, stuck approvals, and pending assignments, then summarizes what needs attention.

Example prompt: “What tasks are overdue across all active workflow runs this week? Group them by assignee.”

No report builder needed. The MCP pulls live data from your workspace and your AI organizes it however you ask.

3. Run a workflow from a conversation

Start a new workflow run, assign it, and fill in form fields without switching tabs.

Example prompt: “Create a new run of the Quarterly Access Review workflow and assign it to the compliance manager, if it’s not already been assigned to someone.”

The AI finds the workflow, creates the run, sets the assignee, and confirms. You stay in the conversation.

4. Find bottlenecks across active runs

When processes stall, the MCP can surface where and why.

Example prompt: “Find the most recent Client Onboarding workflow runs and show me where there are bottlenecks from incomplete tasks.”

The AI pulls the runs, identifies which tasks are incomplete and who they are assigned to, and gives you a clear picture of what is stuck. No clicking through individual runs.

5. Manage data sets with natural language

Data sets in Process Street hold reference data: vendor lists, asset registers, approval matrices, inspection schedules. The MCP gives your AI assistant full read and write access.

Example prompt: “Search our Active Vendors data set for any vendor that hasn’t been reviewed in the past 6 months and mark them as Needs Review.”

Bulk updates that would take 30 minutes of manual clicking happen in one prompt.

6. Edit live workflows without touching the UI

Need to add a field, rename a task, update a logic rule, or wire a new approval gate? Describe the change. The AI resolves the current workflow structure, opens a draft revision, applies the edit, shows you a before-and-after diff, and publishes on your confirmation.

Example prompt: “Add a manager approval task after Submit Report in the Inspection workflow, and rename the Notes field to Internal Notes.”

The MCP handles the revision API mechanics: task ordering, widget schemas, logic rule targeting. You describe what you want changed.

7. Automate compliance evidence collection

Compliance teams use Process Street to enforce policy and collect audit evidence. The MCP extends that into AI-driven automation: scheduled runs, form field pre-population from external systems, and approval chain management.

Example prompt: “Every Monday at 9 AM, check if there are any SOX control test workflow runs from last week that are still incomplete. List them with the control owner and days overdue.”

Pair this with scheduled workflows and outgoing webhooks (both available via the API) for fully automated compliance monitoring.

8. Build conversational interfaces on top of your processes

The MCP turns Process Street into a backend that AI agents can operate on. Teams are building Slack agents that query pending tasks, Power BI dashboards that pull from the API, and custom front-ends that let end users interact with workflows without ever logging into Process Street.

Example: A Slack bot that lets a team ask “What’s pending for my team this week?” and gets a structured answer pulled directly from Process Street runs and tasks.

The MCP server works with any client that supports the protocol: Claude, ChatGPT, Copilot Studio, or your own application.

9. Bulk-assign, reassign, and triage work across your workspace

Operations managers spend too much time reassigning tasks when someone is out sick or a team restructures. The MCP handles bulk operations.

Example prompt: “Reassign all of Sarah’s open tasks across all active runs to Michael, and snooze anything due after Friday.”

The AI reads the My Work queue, filters by assignee, updates each task, and confirms the changes. What used to be an afternoon of clicks becomes one prompt.

10. Connect Process Street to your full tech stack

The 105 API endpoints behind the MCP cover workflows, runs, tasks, pages, data sets, folders, permissions, webhooks, and attachments. That means your AI assistant can build real two-way integrations between Process Street and whatever else you run: CRM, ERP, HRIS, ticketing systems.

Example prompt: “When a new employee is added to our HRIS, create a run of the New Hire Onboarding workflow, set the employee name and start date from the HRIS record, and assign it to their manager.”

The API handles incoming and outgoing webhooks, scheduled workflows, and attachments. The MCP makes all of it accessible through conversation.

Getting started

The MCP server is available on all Process Street plans. Connect it in under two minutes:

• Claude: Open Connectors, search for Process Street, sign in.
• ChatGPT: Settings → Connectors → Add Connector, enter the MCP server URL.
• Copilot Studio: Add a tool, select Model Context Protocol, search for Process Street.

Full setup instructions: process.st/help/docs/mcp-server

API documentation: public-api.process.st/api/v1.1/docs

Frequently asked questions

What is the Process Street MCP server?

The Process Street MCP server is a Model Context Protocol endpoint that wraps the Process Street public API. It lets AI assistants like Claude, ChatGPT, and Microsoft Copilot Studio read and write to your Process Street workspace: creating workflows, managing runs, updating tasks, querying data sets, and controlling permissions.

Which AI platforms support the Process Street MCP?

Claude (claude.ai), ChatGPT (Pro, Team, Business, Enterprise, and Edu plans), and Microsoft Copilot Studio all support the Process Street MCP server. Any application that implements an MCP client can connect.

Do I need to write code to use the MCP?

No. The MCP server connects through your AI platform’s built-in connector or integration settings. Once connected, you interact with Process Street through natural language prompts. No API keys, no scripts, no middleware.

Is the MCP available on all Process Street plans?

Yes. The MCP server and all 105+ API endpoints are available on every Process Street plan.

What permissions does the MCP use?

The MCP server uses your Process Street login credentials. Admins, Builders, and Users can use it, and each person can only interact with the resources they already have permissions to access. The MCP does not bypass any workspace permission settings.

The post 10 Ways to Use the Process Street MCP first appeared on Process Street | Compliance Operations Platform.