User Evaluation is an AI agent-based user research platform with a core goal: to let you ask real customers questions and get evidence-backed answers.
-
Purpose The platform is used for user interviews, product experience feedback, and user behavior research. You pose the research question, and the AI moderator Eva writes an interview guide containing 5 questions and screening criteria in seconds; after you approve it, real, screened participants answer via live voice call. The AI moderator follows up based on your guide and each answer, for example, "When exactly was that?" "Who did you show it to?" You can listen in online and modify the next question. The interview produces a research report within 48 hours, and every conclusion in the report links to the recording clip it corresponds to—if a clip cannot be provided, the platform will not include that conclusion in the report.
-
Main Capabilities
-
AI-moderated interviews: The AI moderator uses your guide during the call and follows up on each answer.
-
A screened and paid participant panel: The platform screens participants from its own panel and schedules them for you; you can also invite your own users with a link. You pay per interview, and recruitment is included.
-
Language adaptation: The AI conducts and transcribes the interview in the participant's language, and you then receive the research report in your own language.
-
Speaker-labeled transcripts: Every transcript has time codes and speaker labels, and is linked to the recording; clicking a line jumps to the corresponding position in the recording.
-
Every conclusion has a citation: Every sentence in the research report links to a file and timestamp.
-
Transcript clips: Select part of a transcript to create a clip, and each clip links to its position in the recording.
-
Report export: Research reports and reports can be exported to PPTX and DOCX, as well as PDF, Markdown, or CSV.
-
Free share links: Share research reports, reports, or files via read-only links; links can be set to expire, and readers do not need an account.
-
Agent session replay: Swarm records every step the AI agent takes in your product, and every issue has a clip from the session replay; you can view the replay to verify the finding.
-
Pre-cost approval: Every operation that incurs a cost first shows the total, and it will not start without approval.
-
API and Webhook: Start research, retrieve research reports, and receive new findings in your own code; you can use the MCP server to connect Claude Code or Cursor.
-
Upload your own recordings: Upload calls, PDFs, and CSVs to the same evidence library and get the same transcripts, labels, and citations as interviews.
-
Swarm agents: AI agents register for your online product as real users, each agent maintains a real account, returns over several weeks, compares plans like a customer, starts a trial, invites teammates, and stops at the paywall. Important findings are confirmed by real people.
-
Unified evidence repository: transcripts, agent sessions, uploaded documents, and recordings go into the same evidence repository.
III. How to use it
- Enter your question: Eva writes a draft guide containing 5 questions and screening criteria, and will not begin any interviews until you approve it.
- Real people answer: the platform screens each participant, and every interview is a live voice call with an AI moderator.
- Read the research report: get it within 48 hours, with every conclusion linked to the clip it came from.
IV. Use cases
- Validate whether a research question holds up: three interviews are usually enough to reveal causes, such as the events that occur before people cancel, stop, or switch to another product. Three interviews are not a sample for deriving percentages; the research report gives a confidence level for each conclusion.
- Product experience testing: Swarm agents operate in the real product, flag issues, and then real people confirm the important findings.
- Incorporate existing calls, documents, and recordings into unified analysis.
V. Limitations and pricing
- Participant sources and pricing: your own people $10 (you send the link); ordinary users from the platform's screened panel $39; professionals screened by role and seniority $99. The price includes all costs such as recruitment. If participant quality is low, mark that interview and get a free replacement participant.
- The first three interviews are free. Interviews start at $10, with no demo call required.
- Recordings are made with participant consent before the call begins; participants are told who commissioned the research. Workspace data is private by default, and data is stored in the EU and the US.
- For other questions, email support@userevaluation.com and a real person will reply.
