Candidate data: name, email, phone, location, LinkedIn URL, pitch and "about" text, resume text, uploaded file names and links, tags, source, ratings and reviews, answers to questions

Recruitment data: jobs, job pipelines and stages, applications, interview kits, scorecards, skills and traits, questions and answers, custom fields

Communication and activity: comments, messages with candidates, activity timelines, meeting details and meeting transcripts

Organization data: employees (name, title, role), departments, divisions, roles, locations

Task data: todos and their assignees

Execute the tool calls that the user's AI assistant makes on their behalf (read, create & modify, delete) in their account

Authenticate and authorize each request

Keep an audit trail of changes made through the MCP server, so the customer can see which user changed what and through which MCP client

Enforce customer's Company Admin permission settings

provide, secure and maintain the MCP service

detect, investigate and prevent unauthorized or abusive use

monitor errors, performance and service reliability

troubleshoot technical issues; and

produce aggregated statistics and usage insights to understand and improve the service

Reads and writes go directly against user's Teamtailor account data, subject to their user permissions and the OAuth scopes granted (read, create & modify, delete).

Every write made through the MCP server is recorded in the customer's audit log with user identity, the connected client's name, user IP address, and a summary of the action.

Teamtailor stores application and infrastructure logs relating to the MCP server, including data points such as tool name, arguments, status and duration, tied to the user, company and connected MCP client and technical error information. Teamtailor acts as data controller when collecting and storing these logs.

The MCP client that the customer's users connects via the integration: tool results are returned to the AI assistant connected (for example, ChatGPT). Once data reaches that MCP client, its handling of data is governed by that provider's own terms and conditions, privacy policy, and by the settings that the customer has with them, including whether or not data is used for training. Review those before connecting.

The customer: Company Admins or other authorised users within the organization can access MCP connection information, permissions and audit records in accordance with the functionality and access controls available in Teamtailor.

Teamtailor's subprocessors: Teamtailor uses the same infrastructure, hosting, security and monitoring providers used to provide the wider Teamtailor service. These providers are listed in Teamtailor's applicable subprocessor documentation.

Authorities and other parties involved in a potential or existing legal proceeding: Teamtailor may only disclose data only as described in the service agreement in place with Teamtailor and as stated in our general Privacy Policy.

Account data (candidates, jobs, applications, etc.) is not stored as a separate dataset by the MCP server. It remains in the customer's Teamtailor account and is retained according to the customer's retention settings and Teamtailor's standard retention rules.

OAuth tokens remain valid until they expire or the connection is revoked by the user or Company Admin.

Audit log events are retained in accordance with the audit-log retention applicable to the customer's Teamtailor account.

Application and infrastructure logs are used for security, troubleshooting, performance monitoring and usage insights and are retained in accordance with Teamtailor's standard retention rules.

Enabling the integration with the customer's chosen AI service provider: connecting an MCP client is opt-in and consent based. The customer authorises it through Teamtailor's OAuth flow and chooses the access level (read, create & modify, delete) per role.

Revoking access: Users can disconnect a MCP client at any time from their Teamtailor account settings, or by asking the Company Admin to revoke the connection. Once the connection has been revoked, the MCP client will no longer be authorized to make further requests to Teamtailor through the integration. Revocation prevents future access through Teamtailor but does not automatically delete data that the AI service provider has already received. The AI provider's retention and deletion practices are governed by its own terms and privacy documentation.

Admin controls: Company Admins control whether MCP access is enabled, and which user roles can use it and at what access level (off, read, create & modify, delete).

User permissions: MCP tools are limited by the permissions associated with the user settings in the Customer's Teamtailor account. An MCP client is not able to access data that a user account is not permitted to access in Teamtailor.

Data-subject rights: For rights concerning candidate, recruitment or other data controlled by a Teamtailor customer, please contact the relevant company directly. Teamtailor will assist the customer in responding to requests in accordance with the applicable Data Processing Agreement.

determining which MCP clients and AI services its users are permitted to connect;

assessing whether the selected service is appropriate for the data concerned;

informing users of and enforcing applicable internal AI, privacy, security and acceptable-use policies

entering into any required agreement directly with the MCP client; and

ensuring that its selection, configuration and use of the AI service complies with the EU AI Act and/or any other laws or regulatory requirements applicable to the customer