Add multiple transcripts to an AI Workspace and ask a question in plain English. Sonix can use the transcripts in the workspace to provide an answer based on the available content.

For example, if you have a collection of interviews, you could ask:

What were the most common concerns mentioned across the interviews?

Sonix can analyze the available sources in the Workspace and provide an answer based on the information found in those sources.

AI Workspace answers include references to the source material. You can use these citations to identify the recording and timestamp where the information came from.

This makes it easier to verify an AI-generated answer by going directly to the relevant point in the recording.

AI Workspaces provide pre-built prompts for common analysis tasks. These can help you quickly:

You can also create your own prompts to ask specific questions or extract particular information from your transcripts.

For example:

What product features were requested by customers? Group the requests by topic and include the relevant timestamps.

Custom prompts can be useful when you need to perform the same type of analysis repeatedly across different transcripts.

Folder-level analysis allows you to analyze a collection of transcripts at once. This can help you identify aggregate insights, recurring themes, and patterns across interviews, meetings, research projects, or other groups of files.

AI can generate summaries of your transcripts to help you quickly understand the key points, decisions, and action items without reviewing the entire recording.

Summaries can be generated in paragraph or bullet-point formats.

AI can automatically organize long transcripts into logical chapters with timestamps. This allows you to quickly navigate to specific topics or sections of a recording.

Thematic analysis helps identify recurring themes and patterns across your content.

This can be particularly useful for research interviews, focus groups, customer feedback, and other qualitative research.

Sentiment analysis helps identify the emotional tone of conversations and changes in sentiment throughout the discussion.

Topic detection identifies subjects discussed in your transcripts and helps organize them so you can find specific discussions more easily.

Entity detection can identify information such as:

Make sure you are signed in to the correct Sonix account and that your subscription supports AI Workspaces. If your account has moved to the Storage Plan, AI Workspaces are currently unavailable; contact Sonix Support if you believe you should have access.

AI Workspace access is currently limited to the Workspace owner, and Admin access does not allow access to another user's AI Workspaces. Admins may still have access to the underlying media, so access should only be granted to users who are authorized to view private or sensitive content.

Make sure the file has been successfully uploaded and fully transcribed before attempting to analyze it. If the transcription is still processing, wait until it is complete and try again.

Try making your question more specific and include instructions about the type or format of response you want. For example, ask the AI to group product improvements by theme and provide relevant timestamps rather than simply asking it to tell you about the interviews.

AI-generated responses should always be reviewed against the original source material. Use the source citation provided with the response to check the relevant recording and timestamp.

Make sure all the transcripts you want to compare or analyze are included in the same AI Workspace or folder before submitting your question.

On new subscription plans, the folder-level AI Analysis option is being phased out as an expected product change. Use an AI Workspace to organize related transcripts and perform AI-powered analysis across them.

For example, if you have 20 customer interviews, you could ask:

What are the three most common complaints across all interviews?

This allows you to identify common patterns without manually reviewing every transcript.

When an AI Workspace response includes a source citation, the citation identifies where the information came from.

Click the citation to access the relevant recording and timestamp. You can then listen to the original recording and verify that the AI-generated response accurately reflects what was said.

We recommend checking the original source whenever the information is important or needs to be cited in a report, publication, legal document, or other high-stakes work.

When you start an analysis, you can select a template for a common analysis task.

Available templates include:

Select the template that best matches the type of information you want to extract. You can also ask a custom question if the available templates do not match your needs.

AI Workspaces provide a way to organize related transcripts and documents and analyze information across multiple sources.

AI Analysis provides AI-powered analysis of transcribed files, including features such as summaries, chapters, thematic analysis, topic detection, entity detection, sentiment analysis, and custom prompts.

On the new subscription plans, the folder-level AI Analysis option is no longer available. This is an expected product change and is not a bug.

Customers on legacy plans may still see the folder-level AI Analysis option, but it is not supported on the new plans.

Instead, use AI Workspaces to organize related transcripts and run AI-powered analysis across them. AI Workspaces are the recommended replacement for managing AI Analysis across related transcripts.

For further assistance, please contact Sonix Support at [email protected] or through the chatbox located in the bottom-right corner of our website.