The Overview page on a study's Analysis tab is the executive summary: one written answer to your research objective, built from every interview in the study. Every number and quote in it links back to the participants behind it, so you can check any finding before you share it.
Open it from Study > Analysis > Overview. The Questions tab next to it breaks the same analysis down question by question.
Tellet starts analysing once a study has enough completed interviews to say something meaningful: 10 completed interviews, or 10% of your max participants if that's higher. Completing the study also starts the analysis, even if you have fewer interviews than that.
Until then, the page shows Waiting for more responses before summarising. Test interviews never count towards this or appear in the analysis.
Once the threshold is reached, the summary takes up to a minute to write. While it's generating, you'll see the live question results on the page, with Generating summary… underneath.
You don't need to rerun anything. As new interviews come in, the numbers in the summary update straight away. If new data changes a finding enough that a sentence no longer reads true (for example, something "most" participants said drops to "about half"), that part of the summary is rewritten automatically a few minutes later.
At the top is a one-sentence headline answering your research objective, followed by the number of participants analysed and a short introduction to who you heard from.
Below the introduction, findings are grouped by research objective. The objectives come from the Research objective field in your study settings: if it describes several goals, each one gets its own block. If it describes one goal, there's a single block.
Each block has:
the objective, restated as a heading
the key finding in one sentence
a short paragraph explaining it
the strongest themes behind it, each with a pull quote from a participant
At the end, the summary calls out views that didn't fit the main themes but are worth knowing about, if there are any.
A clear, specific research objective gives you a clearer summary. See Study settings for how to write one.
Words like most, about half or few are underlined because they're calculated from live data, not written by the AI. Hover over one to see the exact numbers, for example "18 of 24 participants". Click it to open the supporting quotes.
Word | Share of participants |
|---|---|
All | 100% |
Nearly all | 90% or more |
Most | 60% or more |
About half | 40–60% |
Some | 20–40% |
Few | Less than 20% |
Quotes are always the participant's exact words, never paraphrased. Hover over a quote to see the full message it came from, with the quoted part highlighted. Click it to open that participant's transcript at that point in the interview.
The analysis is written in your study's analysis language. To read it in another language, use the Translate menu at the top right of the Analysis tab:
View in: switch between the original language and any languages you've already added.
Add language: translate the summary into a new language. You can add up to 3 languages.
Quotes are translated too. If a quote can't be translated, it's shown in the participant's original language.
Too few responses to analyse: the study doesn't have enough usable answers to summarise yet. Collect more interviews.
Summary could not be generated: click Try again. If it keeps failing, contact us at [email protected].
Failed to load analysis: refresh the page. If the problem continues, contact support.