Reach plc
Guten, a generative AI tool for article tagging and drafting
Reach cut breaking-news publishing time from 9 minutes to 90 seconds with its in-house Guten AI tool
Problem
Reach publishes more than 120 brands, including The Mirror and The Express. Journalists spent time on non-core tasks such as tagging articles and producing first drafts, which slowed publishing of breaking news.
Approach
Reach's data team built Guten on foundation models in Amazon Bedrock. It began as a proof of concept with a handful of journalists at The Mirror and OK Magazine, then spread to Reach's other titles within a year. Guten adds article tags and drafts content for journalists to refine, and it connects to the content management systems. Journalists review all generated copy before it is published.
Results
Takeaway
Reach started with two titles and a small group of journalists and widened the tool once it worked. Human review stayed in place for all generated copy.
Categories
- Industry
- Media and entertainment
- Business function
- Internal work and knowledge management
- Innovation type
- Process innovation
- AI technology
- Generative AI
- Adoption
- Built in-house
- Evidence
- Vendor case study
Sources
Source summaries
An AWS solutions architect and two Reach data leaders describe Guten, a generative AI product built on Amazon Bedrock models. It tags articles and drafts content, and journalists review all output. The post gives the 9 minutes to 90 seconds publishing time and a 1.8 billion page view figure for 2024, and says Guten began two years earlier.
Full details are in the original sources.