MCP is included in the Professional Plan and can be trialed for 30 days, free of charge.
Teamflect's Model Context Protocol (MCP) integration allows your employees to securely connect Teamflect to supported AI assistants and interact with Teamflect directly from those tools.
Once connected, users can ask questions about supported Teamflect information and perform supported actions using natural language.
For example, a user could ask their AI assistant to:
Show their active Teamflect goals
Retrieve reviews that require their attention
Access supported information about their direct reports
Retrieve relevant Teamflect calendar information
Create a goal
Update a goal
Create a task
Create a recognition
This guide explains what the Teamflect MCP integration does, how to enable and manage access, how Teamflect permissions are applied, which AI tools users can connect to, and the capabilities available through MCP.
What Is MCP?
MCP stands for Model Context Protocol. MCP is an open standard that allows AI applications to connect to external tools and data sources.
In simple terms, Teamflect MCP allows users to bring Teamflect into the AI tools they already work in.
Instead of switching between an AI assistant and Teamflect to find information or complete an action, users can ask their AI assistant to interact with supported Teamflect functionality on their behalf.
For example, a manager preparing for a conversation could ask:
"Show me my direct reports and their active Teamflect goals."
Or, after identifying an action item during an AI conversation:
"Create a Teamflect task for me to follow up with Alex about the development action we discussed."
The AI assistant uses the appropriate Teamflect MCP capability to retrieve the information or perform the supported action.
Before You Enable MCP
Before enabling the integration, keep the following requirements in mind:
MCP is available with the Teamflect Professional plan.
A Teamflect admin must enable the integration.
Admins can control who is allowed to use MCP.
Users must have an active Teamflect account.
Users need access to a supported MCP-compatible AI assistant.
Users must authenticate with Teamflect when connecting their AI assistant.
Important: Enabling MCP does not give users additional Teamflect permissions. Teamflect identifies the authenticated user and applies their existing Teamflect access when MCP tools are used.
Step 1: Enable the MCP Integration
To enable MCP:
Go to the Teamflect Admin Center.
Navigate to Integrations.
Locate the MCP integration.
Enable the integration.
Once MCP is enabled, you can determine which users should be allowed to use the integration.
Step 2: Choose Who Can Use MCP
You don't need to make MCP available to every Teamflect user. Use the access settings in the Admin Center > Integrations > MCP to determine who should be able to connect Teamflect to supported AI assistants.
This gives your organization control over the availability of the integration while allowing selected employees to incorporate Teamflect into their AI workflows.
Users who have been granted access will see the MCP option in their Teamflect settings.
They can then retrieve the connection information (URL) required to connect Teamflect to a supported AI assistant by going to 'My Setting' and copying the MCP Server URL.
How Authentication and Permissions Work
MCP does not create a separate permission system in Teamflect.
When a user connects an AI assistant to Teamflect, they authenticate using their Teamflect account since adding the connector to the AI tool redirects them to the Teamflect login page first where they also need to agree and allow access to the AI tool.
Teamflect can therefore identify who is making the request and apply the appropriate Teamflect access.
The information or actions available through MCP depend on two things:
The user's existing Teamflect permissions
The capabilities currently supported by Teamflect MCP
For example:
An employee can retrieve supported information available to their Teamflect account.
A manager may be able to retrieve supported information about their direct reports.
An admin may have broader access depending on their Teamflect admin permissions and the MCP tool being used.
Does enabling MCP give employees additional access?
No. MCP does not override the employee's existing Teamflect permissions.
Does MCP expose all company data to the connected AI assistant?
No. Connecting Teamflect through MCP does not automatically provide unrestricted access to your organization's Teamflect data. Teamflect evaluates requests using the authenticated user's identity and permissions.
Can someone access Teamflect if they obtain the MCP Server URL?
The MCP Server URL alone does not authenticate a user. Users must authenticate with Teamflect when establishing the connection. The URL tells the AI assistant where to connect; Teamflect authentication determines who is connecting and what they're allowed to access.
Where Can Users Connect Teamflect?
Users who have been granted MCP access can connect Teamflect to supported MCP-compatible AI tools.
The initial supported setup includes:
Claude Desktop
Claude Code (CLI)
ChatGPT
Users retrieve their Teamflect MCP Server URL from:
Teamflect > My Settings > MCP
They then add Teamflect as an MCP server or custom connector in their preferred supported AI tool and authenticate with Teamflect.
Option 1: Set Up Teamflect in Claude
Teamflect can be added to Claude as a custom connector using remote MCP.
For Claude Team and Enterprise organizations, an Owner or Primary Owner first adds the Teamflect connector to the organization.
Option 1: Add Teamflect to Your Claude Organization
In Claude, go to:
Organization Settings > Connectors
Select Add, then choose Custom > Web
Enter the Teamflect MCP Server URL.
Complete the configuration and select Add.
Adding Teamflect at the organization level makes the connector available to eligible members. It does not automatically connect every employee's Teamflect account.
Option 2: Have Employees Connect Their Accounts
Employees can then go to:
Customize > Connectors
They locate Teamflect and select Connect (+).
The employee completes Teamflect authentication and authorizes the connection using their own Teamflect account.
This allows Teamflect to apply that employee's existing Teamflect permissions after agreeing and allowing the AI tool to access their information.
Use Teamflect in Claude
Users can enable Teamflect for a conversation through the + menu and Connectors.
Claude can then use the Teamflect tools available through the connection.
Try:
"Who am I in Teamflect?"
Then:
"Show me my active Teamflect goals."
Individual Claude Plans
Users on supported individual Claude plans can add a custom remote MCP connector themselves from: Customize > Connectors > + > Add custom connector
They enter the Teamflect MCP Server URL, add the connector, and complete Teamflect authentication.
Admin consideration: If your organization manages Claude centrally, we recommend following your organization's Claude connector policies rather than asking employees to configure Teamflect independently.
Option 2: Set Up Teamflect in Claude Code
Claude Code connects to Teamflect differently because MCP servers are configured through the command line. Users need their Teamflect MCP Server URL from: Teamflect > My Settings > MCP
Step 1: Add Teamflect
Run:
claude mcp add --transport http teamflect YOUR_TEAMFLECT_MCP_SERVER_URL
Replace YOUR_TEAMFLECT_MCP_SERVER_URL with the URL provided by Teamflect.
Step 2: Authenticate
Open Claude Code and enter: /mcp
Select Teamflect and complete the authentication process.
Claude Code will direct the user to the browser when Teamflect authentication is required.
Step 3: Verify the Connection
Run:
claude mcp list
Teamflect should appear among the configured MCP servers.
Users can then test the connection with:
"Who am I in Teamflect?"
or:
"Show me my active Teamflect goals."
Option 3: Set Up Teamflect in ChatGPT
Connecting Teamflect to ChatGPT works differently from connecting it to Claude.
ChatGPT uses remote MCP servers as custom apps. For organizational ChatGPT workspaces, a ChatGPT workspace admin or owner must first make the Teamflect MCP app available within the workspace before employees can use it.
ChatGPT availability: Full MCP support, including Teamflect tools that create or modify information, is currently available in beta for ChatGPT Business and Enterprise/Edu on the ChatGPT web experience. Availability and functionality are controlled by OpenAI and may change.
Set Up Teamflect in ChatGPT
A ChatGPT workspace admin or owner should:
Open ChatGPT Workspace Settings.
Make sure the workspace allows custom MCP apps/developer mode.
Go to Apps and select Create.
Enter the Teamflect MCP server endpoint and the requested app information.
Configure the required authentication method.
Select Scan Tools.
ChatGPT will scan the Teamflect MCP server and identify the Teamflect tools available through the connection.
Complete the Teamflect authorization process when prompted.
Once the configuration and tool scan are complete, select Create.
The Teamflect app will initially appear as a draft.
Test Teamflect Before Publishing It
Before making Teamflect available to employees, we recommend testing the connection. Open a new ChatGPT conversation and select the draft Teamflect app from the tools menu.
Try a simple request such as:
"Who am I in Teamflect?"
Then test the Teamflect capabilities your organization intends to make available. For example:
"Show me my active Teamflect goals."
"Show me the Teamflect reviews that currently require my attention."
"Create a Teamflect task for me to prepare the quarterly review presentation."
For actions that create or modify information in Teamflect, ChatGPT may ask the user to confirm the action before proceeding.
Publish Teamflect to Your ChatGPT Workspace
Once you've verified the connection and available Teamflect tools, publish the Teamflect app from: Workspace Settings > Apps
Only ChatGPT workspace admins and owners can publish custom apps.
Depending on your ChatGPT workspace plan and configuration, additional controls may also be available to determine who can access the Teamflect app and which actions are allowed.
Once published and made available to the appropriate users, employees can select Teamflect from ChatGPT's apps/tools menu and use it in their conversations.
Important: ChatGPT workspace controls and Teamflect permissions are separate. Making the Teamflect app available in ChatGPT does not give an employee additional Teamflect access. Teamflect continues to authenticate the individual user and apply their existing Teamflect permissions.
For complete step-by-step instructions, share our end-user guide here.
What Can Users Do with Teamflect MCP?
Teamflect MCP capabilities are provided through tools. A tool represents something an AI assistant can ask Teamflect to retrieve or perform.
Users don't need to memorize the names of these tools. In normal use, they can describe what they want to accomplish using natural language, and the AI assistant can use an appropriate available Teamflect tool.
Teamflect MCP tools generally fall into two categories:
1. Read Tools
Read tools allow the connected AI assistant to retrieve supported information from Teamflect. For example:
"Show me my active goals in Teamflect."
"What Teamflect reviews currently require my attention?"
"Who are my direct reports?"
"Show me my upcoming Teamflect calendar items."
Read tools do not give the user additional visibility. The information returned remains subject to the authenticated user's Teamflect access.
2. Write Tools
Write tools allow the connected AI assistant to perform supported actions in Teamflect. For example:
"Create a Teamflect goal to complete our customer onboarding redesign by December 15."
"Create a Teamflect task for me to prepare the Q4 performance review presentation."
"Create a recognition for Sarah for helping the team complete the customer launch ahead of schedule."
Only the write actions currently supported by Teamflect MCP can be performed.
Available Teamflect MCP Tools
The following tools are currently available through Teamflect MCP.
Ask Teamflect Help — Searches Teamflect's help documentation to answer "how do I do this in Teamflect?" questions.
Find Me — Identifies who you are in Teamflect so the assistant knows whose data to work with.
Find a User — Looks up a colleague by name or email to find their Teamflect profile.
Get One-on-One Meetings — Retrieves your one-on-one meetings along with their agendas, talking points, and notes.
Get Calendar — Shows your upcoming Teamflect events, meetings, and deadlines.
Get Check-ins — Retrieves your check-in records and the answers submitted in them.
Get My Direct Reports — Lists the people who report to you.
Get Feedback — Retrieves feedback you have given or received.
Get Goals — Retrieves goals and their current progress for you or your team.
Get Individual Development Plan — Retrieves an employee's development plan, including focus areas and planned actions.
Get Nine Box — Retrieves nine-box placements, showing how employees are mapped on performance and potential.
Get Company Policies — Retrieves the company policy documents stored in Teamflect.
Get Recognitions — Retrieves recognitions given to or received by a person or team.
Get Reviews — Retrieves performance review cycles and the completed review forms within them.
Get Role Profile — Retrieves a role's profile, including its responsibilities, required skills, and expectations.
Get Succession Plan — Retrieves succession plans showing the potential successors identified for key roles.
Get Surveys — Retrieves surveys and their results.
Get Tasks — Retrieves tasks with their owners, due dates, and current status.
Write tools
Close Goal — Marks a goal as complete and closes it.
Confirm Teamflect Draft — Finalizes a draft the assistant has prepared and saves it to Teamflect.
Create Goal — Creates a new goal with its owner, due date, and success measure.
Create Recognition — Sends a recognition to a colleague.
Create Task — Creates a new task and assigns it to someone.
Edit Goal — Updates an existing goal's details, such as its title, description, dates, or owner.
Edit Goal Measurement Type — Changes how a goal's success is measured, for example from a percentage to a numeric target.
Reopen Goal — Reopens a goal that was previously closed.
Update Goal Progress — Records new progress against a goal.
Note: The capabilities available through Teamflect MCP may expand as additional tools are introduced.
Real-World Ways Employees Can Use Teamflect MCP
MCP is most valuable when employees use it to bring Teamflect into workflows already happening inside their AI assistant. Here are some practical examples you can share with employees when introducing the integration.
1. Prepare for a Check-In
A manager could ask: "Show me my direct reports and summarize their active Teamflect goals so I can prepare for my upcoming check-ins." This can help managers bring relevant Teamflect context into their preparation without manually navigating between employee profiles and goals.
2. Check What Needs Attention
A user could ask: "Show me the Teamflect reviews that currently require action from me." This provides a quick way to identify relevant Teamflect actions while working in their AI assistant.
3. Review Current Priorities
A user could ask: "Show me my active Teamflect goals and summarize their current progress." This can provide a quick overview of the user's current goals and priorities.
4. Turn an AI Conversation into a Task
After discussing a project, meeting, or action plan with their AI assistant, a user could ask: "Create a Teamflect task for me to prepare the customer onboarding proposal by October 15." This helps turn a conversation into a tracked Teamflect action without manually recreating the task.
5. Turn a Plan into a Goal
After working through an objective with their AI assistant, a user could ask:
"Create a Teamflect goal to increase customer onboarding completion to 90% by December 31." The supported write capability can turn the agreed objective into a Teamflect goal.
6. Recognize Someone in the Flow of Work
A user could ask: "Create a Teamflect recognition for Sarah for stepping in to support the customer launch and helping the team meet the deadline." This makes it easier to turn positive moments into recognition while they're still top of mind.
Example Prompts to Share with Employees
Employees can use natural language when interacting with Teamflect.
Here are some prompts to help them get started.
1. Understanding My Priorities
"Show me my active Teamflect goals."
"Summarize the current progress of my active Teamflect goals."
"What Teamflect items currently require my attention?"
2. Reviews and Performance
"Show me the Teamflect reviews that require action from me."
"What review-related actions do I currently have in Teamflect?"
3. For Managers
"Who are my direct reports in Teamflect?"
"Show me the active Teamflect goals for my direct reports."
"Summarize my team's active goals so I can prepare for my next check-ins."
4. Creating Actions
"Create a Teamflect task for me to follow up on today's action items by Friday."
"Create a Teamflect goal to complete our onboarding redesign by December 15."
"Create a Teamflect recognition for Alex for supporting the team during the product launch."
Tip: The more context users provide, the easier it is for their AI assistant to prepare an accurate Teamflect action. For write actions, users should review important information such as names, dates, titles, and descriptions before confirming the action.
Before Rolling MCP Out to Employees
Before making MCP broadly available, we recommend:
Decide which employees should have access.
Confirm that your Teamflect organizational hierarchy and user information are up to date.
Let employees know which AI assistants your organization supports or allows.
Explain that their existing Teamflect permissions continue to apply.
Share examples relevant to how your organization uses Teamflect.
Encourage employees to review AI-generated information and proposed write actions before relying on or confirming them.
Share the Teamflect end-user MCP guide.
Frequently Asked Questions
Is MCP available with every Teamflect plan?
No. Teamflect MCP is available with the Professional plan.
Can I control who can use MCP?
Yes. Admins can configure whether MCP is enabled and which users can access the integration.
Does enabling MCP give users more access to Teamflect?
No. A user's existing Teamflect permissions continue to apply.
Can users access another employee's information through MCP?
Only where their existing Teamflect access and the relevant MCP capability permit it.
MCP does not provide unrestricted access to other employees' information.
Does having the MCP Server URL provide access to Teamflect?
No. The user must authenticate with Teamflect when establishing the connection.
Can users make changes in Teamflect through MCP?
Yes, for the write actions currently supported by Teamflect MCP. See the Available Teamflect MCP Tools section above for the current capabilities.
Do users need technical knowledge to use MCP?
No. Once the connection has been configured, users can interact with Teamflect using natural-language requests through their supported AI assistant.
Which AI assistants can employees use?
Teamflect currently provides setup guidance for Claude Desktop, Claude Code, and ChatGPT.
What should I send employees after enabling MCP?
Share the end-user guide. It walks users through finding their MCP Server URL, connecting Teamflect, testing the connection, understanding available capabilities, and getting started with example prompts.