Dashboard
Your dashboard gives you an at-a-glance view of feedback volume and satisfaction, with a date range picker at the top to scope most of it to a specific period.
Key highlights
A short, most-severe-first list (capped at 6) of observations about your responses: things like unresolved negative responses, a page whose satisfaction is trending down, or a submission spike. Each highlight can include a button that jumps straight to a pre-filtered view of Responses.
Highlights aren't all scoped the same way: some reflect your selected date range, others (aging unresolved items, milestones, page-trend comparisons) always look at a fixed period or all-time data regardless of the filter. Those carry the same ⓘ marker described above, right on the individual highlight rather than the whole panel.
Stat cards
Four cards always ignore the date filter, each computed over the trailing 30 days vs. the 30 days before that:
- Satisfaction trend: the change in average satisfaction score between the two periods.
- Published feedbacks: how many responses were collected through each survey's currently published version, compared to the prior period (a response collected under an older, since-replaced draft doesn't count).
- Feedbacks per day: the change in average daily submission volume.
- Best & worst day: the single calendar day (within the last 30 days) with the highest, and lowest, average satisfaction score across everyone who responded that day.
Everything else — scoped to your date filter
The rest of the dashboard does move with the date picker:
- Unique users per day and Total received feedbacks: two per-day charts over the selected range.
- Feedbacks per page: a positive/negative breakdown of responses, grouped by the page they were submitted from.
- Browsers and Operating systems: a table per category showing share of responses and average satisfaction score (1 to 5) for each browser/OS.
- Emotional trendline: a chart of average daily satisfaction across the selected range. Its big number and the chart line are filtered like everything else here, but the small "vs last month" badge next to it is not (it's the same fixed 30-day comparison as the stat cards above, marked with the same ⓘ icon).
AI reports
AI reports, next to the date range at the top of the dashboard, opens reports that AI writes from your survey responses. They are available on premium plans.
Allowing AI analysis
The first time, the dialog asks for permission to send responses to the AI provider. Only a space admin can give it, with Allow AI analysis, and it covers the whole space. Anyone else sees a note asking them to turn to an admin of the space.
The dialog says what is sent: the main open-text answer of each response a report covers (the surveys, period and other filters you choose), with its score. Email addresses and phone numbers typed into an answer are removed first. Depending on the kind of report, the names of the surveys it covers, the names and descriptions of your themes, or the names of the segments it compares are sent too. A segment you don't name is named after its filters, such as a device or a page address. In a summary of several surveys each response is sent with its survey's name; in a comparison of segments, each response is sent with the segments it belongs to. Apart from that, respondents' contact details, IP addresses and other response data are never sent. The details are in the privacy policy and the data processing agreement.
An admin can withdraw the permission at any time with Withdraw permission at the bottom of the dialog. New reports stop at once; the reports already created stay.
Response summary
Pick Response summary, then All surveys or one survey, the period (it starts with the dashboard's date range) and the language the report is written in. All surveys reads the answers of every survey on the site whose responses you can read. The form shows how many responses it reads at most and what the summary costs in credits. Create summary reads the main open-text answer of each response, newest first, and can take up to a minute. A survey with too few answers is refused with the reason, and nothing is charged.
The report has:
- a short summary of what respondents say;
- recurring themes, each with its tone (Positive, Neutral, Negative or Mixed), about how many responses mention it, and quotes copied word for word from the answers. The button next to a quote opens that response in a new tab;
- what people like, and what they dislike or ask for;
- what to do next: actions that follow from the answers.
Quotes are checked against the answers that were sent, so a masked email address or phone number shows in them as [email] or [number]. When more responses match than the summary reads, the report says it read only the newest ones. The report is written by AI: check important conclusions against the responses themselves.
Credits are charged only once the summary is ready. They are listed on the Credits page as "AI summary of responses".
Themes and tone per response
Themes and tone per response tags each response with its tone (Positive, Neutral, Negative or Mixed) and with up to three of your website’s own themes. The tags are written onto the responses: each response’s page shows them (see Response detail), and the run’s report counts them.
The themes are the website’s own list, up to 20. Everyone who opens the dialog sees the list; only a space admin can change it. Add theme takes a name and, if you like, a sentence on what the theme covers, which helps the AI tell similar themes apart. Under Suggestions, common themes such as price, delivery or customer support are added in one click. Until someone edits such a theme, everyone sees it in their own language, and its suggestion is not offered again in any language; once edited, it shows as written. A renamed theme stays on the responses already tagged with it. A deleted theme also comes off every response tagged with it.
Pick All surveys or one survey and the period (it starts with the dashboard’s date range). Tag responses reads the main open-text answer of the newest responses with text that no run has tagged yet, at most 500 a run, and can take up to a minute. A tagged response keeps its tags, so the next run tags newer responses and the ones an earlier run left. When there is nothing left to tag, the run is refused and nothing is charged.
The form shows what a run costs in credits, which are charged per block of responses tagged; a part of a block costs as much as a whole one. A response the AI leaves untagged costs nothing. With fewer credits left than a full run needs, the run reads only as many responses as the credits cover. Credits are listed on the Credits page as “AI tagging of responses”.
The report shows how many responses were tagged, how they split by tone, and for each theme how many responses are about it and in what tone. Click a theme to see the start of each of those responses, newest first, with its tone and a button that opens the whole response in a new tab; a masked email address or phone number shows in it as [email] or [number]. The responses no theme fits open the same way. The report also says how many responses with text in the period are still untagged. The tags are written by AI: check important conclusions against the responses themselves.
Segment comparison
Compare segments sets two to five groups of responses side by side, such as mobile and desktop, or this month and last. A segment takes the filters of the Responses page: surveys, period, satisfaction score, device and, under More filters, OS, browser, form type and page. A filter left empty does not narrow the segment, and segments may overlap. The form opens on two segments, mobile and desktop, over the dashboard's date range.
A segment's name is made from its filters until you type your own. The AI tells segments apart by name, so each needs a different one. Add segment adds another, up to five. The form shows how many responses each segment reads at most and what the comparison costs: credits are charged per segment. Compare segments reads the main open-text answer of each response in each segment, newest first, with the same share for every segment. A segment with too few answers is refused by its name, and nothing is charged.
The comparison has:
- each segment's numbers, which Voxtrics counts itself: how many responses with text matched, how many were read and their average score;
- a short summary of how the segments differ and what they share;
- themes, each with a row per segment: the theme's tone in that segment, or Not mentioned, and quotes from that segment's answers. The button next to a quote opens that response in a new tab;
- what sets each segment apart;
- what to do next, naming the segment an action concerns.
Credits are listed on the Credits page as "AI comparison of segments".
Recent reports
Every report stays under Recent reports in the dialog, for everyone in the space who can read the responses of every survey it drew on. A report can be deleted by whoever created it or by a space admin. Deleting the report of a tagging run leaves the tags on the responses. The other report kinds in the dialog are marked Coming later.
Download PDF in an open report lays it out as slides in the language it was written in and opens the print window, where you pick “Save as PDF” as the printer.
The slides start with an overview of the key numbers, such as how many responses were read. A summary’s or a tagging report’s slides show each theme’s share of the responses; a comparison’s slides list the filters that make up each segment.