Livstick Insight · Feature
Themes with sources. Every claim opens its proof.
Insight reads a whole set of messages and tells you what comes back. Each theme stays linked to the messages behind it, so you can go and watch them.
Mentioned in 38 messages, mostly negative.



- A summary
- Up to six themes
- Each theme linked to its messages
- Answers to your questions
- Counts computed by Livstick
- One click from the video
How it works
From a pile of messages to something you can act on.
-


+209
Step 1Choose a set of messages
A campaign, a period, a product. Every message in it has already been analysed with your grid.
-
Step 2
Insight writes the synthesis
A summary, up to six themes, and the answers to the questions you set in the grid.
-
0:42 “Lovely welcome, once we got to the desk.”
Step 3You open the sources
Click a theme to see the messages behind it, and watch the customers who said it.
A synthesis, in use
Six themes on the left. The customers who said it on the right.
Pick a theme and the panel fills with its source messages: who, what they said, and where in the video.
Summary
Guests come back for the breakfast and the welcome. The wait at check-in is the irritant named most often, mainly on Friday evenings. Light sleepers ask for the garden side.
Counts are computed by Livstick, not by the AI.
Theme 2 · sources
Wait at check-in
- “Lovely welcome, once we got to the desk.”
- “Twenty minutes in the lobby after a long day.”
- “With two kids, the queue felt long.”
Your questions
You ask in the grid. The synthesis answers.
Next to the themes, the synthesis answers the questions you set in your grid. The answer is written from the set of messages, and points back to them.
- Your wording. The questions are the ones you wrote in the grid.
- One answer per question, for the set of messages you chose.
- Never detached from the sources. The messages stay one click away.
Breakfast comes first, far ahead of the rest. Guests name the fresh bread and the time they were given to enjoy it.
Based on 64 messagesThe wait at check-in. It is the irritant most often named, and it colours stays that were otherwise positive.
Based on 38 messagesNumbers you can trust
The AI reads. Livstick counts.
A language model is good at understanding what someone says. It is not what you want adding things up. So the two jobs are kept apart.
- Léa’s message Wait at check-in
- Maya’s message Wait at check-in
- Sam’s message Wait at check-in
- Another guest’s message Breakfast
Plain arithmetic on the messages. The number is not written by a model.
- messages linked to Wait at check-in 38


A theme is a claim. The messages behind it are the proof.
Why it matters
A summary nobody can check is just an opinion.
-
Listening to theme 2
-
You can show your work
When someone asks “who says so?”, you open the theme and play the customers who do.
-
Six themes, not sixty
A synthesis holds up to six themes. It forces a priority instead of a list.
-
The words stay theirs
You go from the summary to the video in one click, and hear how it was really said.
-
Numbers that hold
Counts are computed by Livstick, so the same set of messages gives the same totals.
One synthesis, every source.
Whatever door a message came through, it can sit behind a theme.
- Stories
- Video messages inside gifts
- Referral videos
- Voice interviews · preview
- Your own imported media
Questions
Good to know.
-
What is in a synthesis?
A summary, up to six themes, and the answers to the questions you set in your grid.
-
How many themes can there be?
Up to six. Enough to see what matters, few enough to remember.
-
Can I check where a theme comes from?
Yes. Each theme is linked to the messages behind it, so you can open them and watch.
-
Who computes the numbers?
Livstick does. The AI reads the messages; the counts are computed by Livstick, not by the AI.
-
Which messages can a synthesis cover?
A set of messages from any Insight source: Stories, the video messages inside gifts, Referral videos, Voice interviews (preview) and the media you import.