GermanyMedia and entertainment2023 – ongoingProSiebenSat.1 Media
In-house AI Products team automation tools for video, transcription and ad targeting
ProSiebenSat.1 automated news clips, video tagging and transcription with five in-house AI tools
Problem
Editors spent hours documenting and transcribing raw footage. Producing separate regional news clips by hand was not practical.
Approach
The company's AI Products team built five tools. Digital Buzzroom makes automated news video clips. Video mining generates keyword tags for the media archive. Automated transcription of raw footage was piloted with one editorial team. A machine learning model classifies TV viewers for addressable TV ads, and MediaLytics links TV reach to website visits.
Results
Takeaway
Most of the five tools targeted routine editorial and archive work, where a pilot with one team was enough to test them.
Categories
- Size
- Large enterprise
- Industry
- Media and entertainment
- Business function
- Production and qualityMarketing and sales
- Innovation type
- Process innovation
- AI technology
- Predictive analytics and MLComputer vision
- Adoption
- Built in-house
- Evidence
- Company's own claim
Sources
Source summaries
The company story presents five automation products from its AI Products team. It reports 16 automated COVID-19 clips a day and says automated transcription runs about as long as the video, against 3 to 6 hours manually per hour of speech. Transcription was a pilot with one editorial team.
Full details are in the original sources.
