Skip to main content
Question

AI harmony for old content

  • September 16, 2025
  • 1 reply
  • 93 views

While Docebo’s current search functionality leaves much to be desired, having AI Harmony search across all of our platform content would be transformative for our organization. However, if Harmony’s search capabilities are limited to content only from 2025 onward, the value to us is minimal, as our older content holds significant importance.

Is it possible to extend Harmony’s search functionality to include all existing and historical content on our platform, rather than just new material? Access to our complete knowledge base is critical for both ongoing training and knowledge retention.

Additionally, does your platform offer robust support for mass-transcribing videos? 

1 reply

  • Influencer I
  • September 22, 2025

Hi, I know this is not a complete answer, but it may be useful for now.

As per the Harmony search documentation, it states that old SCORM content will not be processed “in the initial release” and “Pre-existing materials will start to be processed at a later stage”. So old content should also be processed, although not yet and we don’t know when.

For other types of content, there are other criteria that define if the content will get processed or not:

will be processed if they satisfy at least one of the following conditions: 
- Training materials or assets: Created from January 2025 onward 
- Training materials or assets: Have been played from January 2024 onwards
- only for Training materials: Are contained within a course or learning plan completed from January 2025 onward

So technically you can force the “processing” by playing the content as a user. As i haven’t tested this, i would recommend to play the entire course, i.e. go through each chapter, since we don’t know if the “processing” is per chapter or per course. Also i don’t know how quickly Harmony would be ready after this, so patience is needed to see the results, but i guess this can also be tested.

Regarding your second question, i’m not sure what “robust support for mass-transcribing videos” means, but in the context of Harmony, in the same documentation article they state that the AI gets its information from the video transcription, without asking you as an admin to upload one, so Docebo must transcribe every video that meets the criteria and further to analyze it for Harmony to be able to provide answers from it.