What does the Overview tab show?
Project-level statistics (completion rate, response count, average duration), an AI-generated executive summary grouped by research objective (each with a headline finding and supporting themes), key themes identified across all conversations, cross-question patterns that recur across multiple questions, strategic recommendations, notable visuals (AI-selected images respondents shared), Voice of the Customer quotes with links to source transcripts, a sample demographic profile of the cohort, and per-question summaries. Completion rate is the share of participants who passed screening and went on to complete the interview; a tooltip next to the metric explains that a low rate can mean the interview is too long or too hard to finish.
What does the green Run button do?
Run starts AI analysis of the project's interview responses and generates the summary, themes, and insights shown in the Overview tab. Once analysis exists, the button becomes Re-run; clicking it reprocesses every conversation from scratch rather than only the new ones.
What does the Questions tab show?
Per-question analysis including an AI-generated summary, AI-suggested categories with response distribution, representative quotes per category, a detailed insights view, and an answers view showing actual participant responses grouped by category. You can toggle between tally (counts) and percentage views, and sort categories ascending or descending.
What are categories and how do they work?
Categories organise responses into labelled groups:
- Single mode: each response belongs to exactly one category (mutually exclusive, totals add to 100%).
- Multiple mode: each response can belong to several categories (totals can exceed 100%). Minimum 2 categories, maximum 30.
Categories can be AI-suggested (the AI analyses responses and generates labels and descriptions) or manually created by you. When you edit and save categories, the AI automatically re-reads all responses and re-assigns them based on the updated descriptions.
What does the Transcripts tab show?
All conversations in three views: Completed (finished interviews), Incomplete (interviews that were started but not finished), and Screened Out (participants who failed screening). Each conversation shows the full message-by-message transcript, timestamps, participant metadata, AI-generated summary, and category assignments. You can sort, filter, translate, delete, and export from this tab.
What is the Research Assistant?
A conversational AI agent that has access to all your project data: transcripts, analysis pages, comparisons, categories, and statistics. You ask questions in natural language and it responds with cited answers (e.g. [R7] links back to the source transcript), structured artifacts (quote collections, data tables, theme summaries, bar charts, markdown reports), and multi-turn conversation context. It works in 64+ languages. You can save threads, share them read-only with teammates, and export sessions as Markdown documents. It sits on the Results page next to the Compare tab, and only appears once you change the project status from Live to Completed.
What kinds of questions can I ask the Research Assistant?
First-pass questions (fast): \"What are the main themes?\", \"Give me an executive summary\", \"What did respondents say about question 3?\", \"How many women mentioned quality?\". Follow-up exploration: \"Find me more surprising responses\", \"Show me specific quotes that support theme 1\", \"Who disagreed?\", \"Dig deeper into pricing\". Respondent-level: \"Show me respondent 5's full interview\". Reports: \"Generate an executive summary with key quotes for a client presentation\".
What is the Compare tab?
The Compare tab generates side-by-side reports between two audience filters. It produces an executive narrative leading with the most important difference, a statistical overview with deltas, sample demographic profiles, shared themes with notes on how each manifests differently, patterns unique to each segment, and a per-question breakdown. Reports can be exported as Markdown. Comparisons are saved so you can return to them later.
How do I filter results?
Go to the Results tab > Overview (or Questions or Transcripts) > Select Filter or Add Filter. Filters can be based on screening question answers, demographic variables from the panel, or metadata passed via URL parameters. Filters apply across Overview, Questions, Transcripts, and Exports simultaneously.
How quickly do I get results?
Analysis updates in real time as responses come in throughout fieldwork. There is no need to wait until the end of fieldwork to start reviewing insights.
Can I delete individual conversations?
Yes. Select conversations in the Transcripts tab, click Delete, and confirm. Deletion is permanent and cannot be undone. Requires Workspace Owner or Editor role. Re-run analysis from the Overview tab after deleting completed conversations.
Deleting a conversation does not automatically remove its quotes from the Overview tab's "In Their Own Words" section. Those quotes are generated during analysis, so they stay in place until you re-run analysis, which regenerates the quote selection from the current set of conversations. If a quote from a deleted response is still showing after a re-run, contact support with the project name and the participant's conversation ID so the team can investigate.
Do you provide sentiment analysis?
Yes. Tellet automatically analyses sentiment across responses, helping you identify emotional patterns and reactions at scale.
How does the AI decide on themes and categories?
The AI analyses response patterns across all conversations and identifies recurring language and sentiment clusters. It generates category labels, descriptions, and logical groupings automatically. You retain full control to review, edit, rename, and reorganise categories at any time.