AI assistant integration
Prepare a one-on-one from the assistant you already use
ModoPersona exposes a bounded Model Context Protocol server for managers who want to prepare from ChatGPT, Claude, Gemini, or another compatible client. The assistant can search and retrieve the manager's own confirmed context, prepare for one direct report, and retrieve bounded history without gaining a write path into the workspace.
Private workspace for one manager. No employee accounts required.
Illustrative product example
Example assistant request
Prepare my one-on-one with Priya
Read-only MCP · manager-authorized scope
- One open manager commitment with its source
- Recent confirmed context relevant to the next conversation
- Older context returned only through an explicit history request
The job to be done
Manager context should travel without weakening its trust boundary
Copying employee context into a general chat loses provenance and creates another uncontrolled record. The MCP connection keeps authentication, tenant isolation, source links, and the distinction between current and historical context inside ModoPersona while letting the manager choose the assistant interface.
How ModoPersona handles it
A focused workflow, with the source still visible
Connect through OAuth
The client discovers ModoPersona's authorization metadata, uses PKCE, and asks the manager to approve the narrow modopersona:prepare scope. Revoking the grant stops future calls.
Use four read-only tools
Search, fetch, prepare_one_on_one, and retrieve_preparation_history return bounded, evidence-linked manager context. None of the tools can create, edit, confirm, or delete workspace data.
Keep judgment with the manager
The assistant presents source-visible preparation. The manager decides what is accurate, what deserves a question, and what becomes confirmed evidence inside ModoPersona.
See the workspace
Context stays useful because you can inspect it
The walkthrough uses synthetic data to show preparation, person-specific context, evidence, and manager commitments in one workspace.
An assistant interface, not an employment decision engine
The MCP server has no write tools and does not score employees, compare people, infer performance, or make employment decisions. Restricted employees retain their restriction boundary, and every tool call is scoped to the authenticated manager's tenant.
Questions managers ask
- Which assistants can connect to ModoPersona?
- The server uses public HTTPS Streamable HTTP and a cross-provider compatibility profile tested for ChatGPT, Claude, and Gemini connection patterns. Individual clients may require their own account or connector configuration.
- Can an assistant change my ModoPersona data?
- No. The exposed tools are read-only. Confirming evidence, editing notes, and changing employee context remain deliberate actions inside ModoPersona.
Continue exploring ModoPersona
- Use casesBring a useful question to your next one-on-onePrepare one-on-ones from past agreements, recent notes, and your own promises. A worked example and a ten-minute preparation guide.
- Use casesManager notes that return when they are usefulKeep private, person-specific manager notes with dates and context, then retrieve them when preparing for a one-on-one.
- Use casesPrepare performance reviews without review-season archaeologyOrganize approved, dated evidence across the review period while keeping the final performance judgment with the manager.
Prepare for one better conversation
Start with one direct report, bring in the context you trust, and see what returns before the next one-on-one.
Start a 14-day trial