Skip to main content
The Research Agent lets anyone with access to a study explore its material in plain language. Ask a new business question, compare participant groups, test a report conclusion or prepare a stakeholder deliverable without losing the link to source evidence. Bring in existing research, including reports, presentations and articles, to explore it alongside your Verso studies. Attach the relevant documents and select the studies you want to draw on.

​
Get a useful answer

1

Choose the scope

Open the agent inside a study, or start from the organization home and select several accessible studies.
2

Ask a focused question

Name the issue, participant group and output you need. For example: “Compare how first-time and repeat buyers describe price. Cite the answers used.”
3

Verify the response

Open citations for the claims you plan to use. Check the original answer, participant context and question that elicited it. Every quote is an exact excerpt of its answer, cuts marked with ”…”, and carries a citation that opens that answer; a quote the agent cannot cite is not shown. Ask for quotes supporting a point and the response opens with how many participants back it, then gives one quote per supporting interview. When you ask for participant videos or recordings, the response lists a short description and citation for each matching answer with a recording, without repeating its transcript; open a citation to play it in the drawer. Switch between Interview and Synthesis when you need the wider summary; the recording remains available in both views, and a citation inside the synthesis jumps back to that answer in the interview.
4

Refine and share

Ask a follow-up to narrow the cohort, examine disagreement or change the format. Request a table, synthesis or available presentation export when ready. Use Export at the top of the conversation to take the whole exchange into Word or Markdown.

​
Questions worth asking

The agent’s home suggests a few questions to start from. When questions have been prepared for the study, or for your organization, it shows a random few of those and selecting one sends it; otherwise a general suggestion fills the composer for you to adapt.
  • What reasons do participants give for abandoning setup, including views outside the main pattern?
  • Where do customers in two markets agree and differ on trust?
  • What supports the report finding about price sensitivity, and what challenges it?
  • Which participants disagree with the dominant reaction to the concept, and why?
  • Where did frustration or hesitation come through on the pricing question, and over which words? (Studies with emotional analysis.)
  • Can you draft a one-page brief with three findings, implications and evidence to review?
When you ask how many participants said something that no code or form question covers, an editor’s agent codes it across every interview and answers with exact counts, which takes about a minute. The new dimension is added to the codebook, where you can review, refine or delete it. The agent says which questions each count covers. When an existing code counts a topic across several questions and you ask for it on one of them (“only when they were asked which factors matter most”), the agent codes the answers to that question alone, its follow-ups included, and gives that figure beside the broader one. The agent works within your access and the product features available for each study. Its responses are generated and may miss relevant material, so verify important claims and counts. If fieldwork continues, rerun the question before treating an earlier answer as current. Use Reading a report for the prepared synthesis, Question analysis for a known question, or Interviews and transcripts for full participant context.