Accessing the import feature
If you are on the Yabble home page, click the ‘Summarize your data’ button:
Alternatively, you can start from the Projects page, then select ‘New project’ > ‘Summarize’.
Tips on uploading your data for Summarize
Are you ready to import your long form qualitative data, such as in-depth video or audio interviews, focus group transcripts, call transcripts or any other form of transcript data?
Preparing your files for successful import
Context. The context is the most important part of the upload – it is key and helps the AI understand the data it's analyzing. Please note this does influence your Summarize output so we recommend spending some time phrasing the context. We’ve included some tips below on crafting your context.
Data format .mp4, .mp3, .docx or .txt. Yabble supports the upload of transcripts in .docx or .txt formats, as well as the option to upload audio and video files for transcription and translation natively in the platform.
Remove files from zip folder. Check that your file(s) have been saved to a folder on your computer and aren’t collectively saved in a zip folder.
Remove tables and pictures from your transcript. Ensure that your transcript file doesn’t include any tables or pictures. The system will not be able to process the file and you will see an error message if there is a table hidden in the file.
Headers & Footers. You can keep header and footer information in the file.
If your file fails to import or you see an error message, see our Summarize FAQs for troubleshooting.
Import your file/s
Follow the prompts to import your files to Summarize.
Start by defining your Project details – project name and industry. Make sure you name your project something that you will be able to easily identify when searching the ‘Projects’ page in your Yabble account.
Add your summary name and context – providing the AI with a context for our Summarize tool is the most important part of the upload, and pretty much the only manual work you need to do before you get your summary from Yabble. Your context is key – it helps the AI understand the data it's analyzing and can affect your Summarize output. We encourage you to take the time to craft a detailed, targeted context.
How to write a strong context for summarize
When you’re uploading your data in Summarize, you’ll see a box asking you to provide context.
Your context helps the AI understand the data it's analyzing. It's extremely important and will affect your Summarize output, so we encourage you to take the time to craft a detailed, targeted context.
Context should be written in paragraph form and can be up to 750 characters in length. It should include:
The type of long-form data you're uploading (e.g. interview transcript, survey, podcast transcript),
The topic of the data (e.g. skincare and beauty products), and
The purpose or intent of the data (e.g. the questions it's asking).
Some tips for writing a strong context:
Be as pointy as possible. Tell the AI exactly what your data is about so it knows exactly what it’s analyzing.
Use phrases like “The intent of the survey is to...," “The goal of the interview is to...," or "The interview asks questions such as..."
Use names wherever possible. If it’s an interview for Google, say that.
A few examples:
This is an interview done on behalf of the national postal service. It contains roughly 4 hours of content. The intent is to understand what’s going on in New Zealanders’ lives at the moment: what are the priorities, what’s not such a priority, and how people are feeling in general.
This is a survey about skincare and beauty products. The respondents are influencers in the beauty industry with 10,000 or more followers on TikTok or Instagram. The intent of the survey is to determine which beauty brands are most exciting to influencers right now, as well as the trends they think will become popular over the coming year.
This is an episode of a podcast with female investors. The goal of the podcast is to uncover how these investors are making decisions about where to invest their money. It’s also designed to find out what’s important to these investors, what they’re cautious of, and what ultimately drives their decision to invest in one institution over another.
Select the files you want to summarize
Once you’ve completed the above steps, select the files you’re summarizing from the table, review and complete! It can take 15 – 30 minutes for your summarization to be complete, and up to an hour if you have uploaded video and audio files. Once it’s finished, view it from your Projects view in the platform.
Uploading Multiple Files
You can upload files to an existing project and use a new context or re-use the same context.
Navigate to Projects
Select the relevant project
Enter your import name and choose the files you want to add.
Confirm our privacy policy and hit Next
When adding additional files and re-using context, your overall summary will be updated to include the newly added file/s. If you create a new context, a new overall summary will be generated for the new file/s.
To re-use your context and update your overall summary in the same project
Navigate to Projects
Select the relevant project
Click on your previously uploaded summary, and click ‘Add more files to summary’
Choose the files you want to add – the project details and summary details will remain constant
Confirm our privacy policy and hit 'Upload and continue’
Confirm the files you’re importing, review and complete. Your new summary will be available in 15-30 minutes.
To create a new context in the same project
Navigate to Projects
Select the relevant project
Click on “+ New summary” on the right to create a new overall summary, with new context
Enter your new summary name, context and choose the files you want to add (following the same steps as earlier).
Confirm our privacy policy and hit 'Create summary’
Confirm the files you’re importing, review and complete. Your new summary will be available in 15-30 minutes.
Now that you have imported your data to Summarize, continue to the next module to learn how to interpret your Summarized data.