Explore analyzes insights across all of your recordings in one place. Instead of reading files one at a time, you see the patterns, trends, and comparisons across your whole library or a single folder.
What you can compare
- Keywords: which terms appear most often across recordings
- Sentiment: how sentiment moves over time or across groups
- Topics: the themes that recur in your content
- Speakers: speaking patterns across multiple recordings
- Entities: people, organizations, and locations mentioned across files
This is how you answer questions that no single file can: the themes running through 50 interview transcripts, the sentiment trend across support calls, the objections that come up most in sales calls, the topics your podcast keeps returning to.
Search across every recording
The Search tab on Explore looks for a word, phrase or topic across everything in scope, and charts how often it comes up and how sentiment moves around it. The box reads “Search for a word, phrase, or topic…”.
A search matches the file name, the description, the transcript, and the values in your
custom fields, so a phone number, an order reference or an address you
captured as a field finds the recording it belongs to. Matching ignores case and looks for the text
anywhere in the value, so part of a number finds the whole one. It matches the value as it is
stored, so a number stored as +447700900123 is found by 447700900123 but not by 07700 900123.
Filter down to what matters
Filters narrow the set of files behind every chart on the page:
- Folders: analyze one project folder
- Date range: look at a period of time
- Speakers: focus on specific speakers
- Sentiment: keep positive, negative, or neutral files
- Tags: group by your own tags
- Categories: filter by insight category
Combine filters with AND and OR conditions to include and exclude data until the view is exactly what you want, then save the filter so you can return to it without rebuilding it.
Read the charts
Explore shows your data as charts and tables. You can also ask AI Chat for a visualization of your own:
- “Create a pie chart of sentiment distribution”
- “Show a bar chart of top keywords this month”
- “Generate a doughnut chart of speaker distribution”
Analyze a large set of files
To go from a pile of raw files to aggregate insights:
- Upload: bring in your audio, video, and text files, or import a CSV of written responses. Check supported formats first, and make sure you have enough transcription hours for the batch.
- Organize: move the files into a single folder.
- Open the folder insights: the folder’s Insights tab aggregates every file in it, with a word cloud of frequent terms, the positive and negative balance across the set, and top mentions of brands, people, and locations.
- Ask across the set: select all the files and run one AI Chat prompt, such as “What is the main pain point?”, to get an answer for every file.
Analytics for a large folder take a moment to generate the first time.
Look at one day, or one day of the week
Click the calendar icon in the top right of Explore or your dashboard and pick a single day, for example October 13 to October 13. Every chart on the page redraws for that day alone, so you can see the volume, sentiment, and keywords it produced.
There is no filter for a repeating day such as “every Monday”. To study that, select all files in
Explore, click Export CSV, open the file in a spreadsheet, and group the rows with a formula
like =WEEKDAY(Date). If you need it as a live filter instead, tag files by their creation day
through an automation and then filter on that tag.
Day boundaries follow your account timezone, so set it to your local time before you read a single-day view.
Evaluating Speak AI for a team? Book a demo and see your own recordings analyzed.
Related: Insights · Categories