Cross-Study Customer Research: Querying Past Research via MCP

Research compounds when a new decision can use evidence from earlier work. An AI agent can help compare findings, but its conclusion is only as complete as the studies and interviews it actually retrieved.

This guide covers cross-study work with the current User Intuition MCP tools and distinguishes it from the dashboard’s Intelligence Hub search.

Research that compounds instead of resets

Before commissioning new interviews, identify what your team already learned, who participated, and when. Reusing old evidence can save fieldwork; it can also mislead if the audience or product has changed. Treat relevance and freshness as part of the analysis.

What the current MCP surface supports

Agents can search findings and participant responses across authorized studies, then retrieve the underlying reports and interviews. Results preserve study context and source links; the calling agent interprets the evidence. Search coverage is explicit, and retrieving evidence does not launch research.

generate_report generates a report for a selected study. It is not a cross-study search or presentation-generation operation. An agent may write a comparison in its own context after retrieving multiple reports, but that is agent synthesis over selected sources.

A source-aware comparison workflow

1. Use list_studies with its documented pagination to identify candidate studies.
2. Retrieve each study with get_study to check its question, audience, and saved plan.
3. Retrieve its report with get_study_report; check interview count and stale status.
4. Use list_interviews and get_interview to verify the evidence behind material claims.
5. Write a comparison with study IDs, time periods, audience differences, and supporting references.
6. State what was not retrieved. A gap in the selected sources is not proof that the entire research library lacks an answer.

Four questions an agent can investigate

Have pricing objections changed? Compare similarly recruited cohorts and distinguish an explicit price complaint from an inferred value concern.

Did onboarding improvements address the earlier problem? Compare interviews before and after the change, allowing for differences in user experience and study questions.

Which messaging themes recur? Retrieve the relevant concept-test reports and inspect the underlying reactions. Keep repeated themes separate from numerical preference claims.

What should the next study investigate? Collect unanswered questions and contradictions. Use those to narrow the next brief instead of asking a broad question that repeats prior work.

How the Intelligence Hub fits

The Customer Intelligence Hub supports research-library exploration in the dashboard. API and MCP workflows can search findings and participant responses across authorized studies, then retrieve the relevant reports and interviews. Describe the search coverage and the sources actually inspected.

An evidence-backed agent answer should make it possible to check the source study and interview. It should also distinguish a participant’s statement, a recurring pattern across retrieved interviews, and the agent’s own interpretation.

Get started

Connect through the MCP setup guide, select two relevant studies, and ask your agent to compare them with source references and explicit coverage limits. See agentic research for creating new research when the existing evidence is insufficient.