Thanks to the AI solution, it is possible for the Funke group to process and analyse content in near real time in a way that wasn't possible before. The platform and its reusable components allow Funke to quickly test and implement new content ideas or products in record time.
Funke Mediengruppe is one of the largest publishing houses in Germany, with a focus on regional media, magazines and digital. The group issues twelve daily newspapers as well as more than 50 magazines - in print and digitally.
As a digital frontrunner, Funke embraces the importance of investing in the right innovation, and sees Artificial Intelligence and Cloud Technology as key steps to accelerate its journey.
The media industry is going through significant change. New capabilities and solutions are needed to keep up with the rapidly developing environment, changing customer demands, and competitive pressure.
Speed-to-publishing and efficiency in production are getting ever more important, as is providing relevant and personalized content to customers. In order to keep and expand a competitive advantage in this environment, it is fundamental to have the capability to rapidly analyse and understand large amounts of content.
Funke Mediengruppe started to collaborate with ML6 to explore how to leverage intelligent technology for working with content, in order to enable new ideas and products - convinced that current and future market challenges could be solved through the smart use of Machine Learning and Natural Language Processing.
Funke and ML6, in very close collaboration, built an AI-driven content platform. The platform is composed of different capabilities to understand and curate content, which serves as the basis for many future use cases and new product ideas.
The goal of the platform is to analyse and understand incoming content streams from different sources with Machine Learning, allowing for near real time processing and curation of large amounts of information automatically. The platform can process various content types, such as text and images. State of the art NLP and Machine Learning knowledge has been used to build the different capabilities of the platform, such as keyword extraction, topic clustering, content matching as well as text generation and summarization.
The development of each capability was strongly focused on reusability, making sure that each component can be leveraged later on. This was particularly important, as the platform idea was specifically aiming at enabling future potential use cases which rely on content understanding, syndication and generation.
Both parties complemented each other during the development of the platform - the passion for news, irreplaceable business knowledge and ideas from the Funke team and the technology expertise from ML6 combined allowed us to build a relevant, robust and future-proof solution.
The platform and its services are built on Google Cloud Platform and is integrated to the existing system landscape supporting the content lifecycle.
Special attention has also been paid to Security, leveraging best practices and using advanced security capabilities on GCP, resulting in a highly secure and reliable system.
As a result, it will be possible to process and analyse content in near real time in a way that wasn’t possible before. The platform and its reusable components will allow Funke to quickly test and implement new content ideas or products in record time.
The creation of the AI solution was done in a gradual way, with Funke and ML6 working closely together in every step of the way to make sure Funke’s industry skills and ML6’s AI expertise were fully integrated and leveraged. A strong focus has also been placed on knowledge transfer and upskilling the Funke team on AI and MLOps topics.
Starting with an ideation workshop, the domain/business knowledge of Funke was combined with the ML knowledge of ML6 to define a joint roadmap for the platform. Working in an iterative way, different components of the platform were developed and scaled up to ML in Production projects. The project was run using the Scrum methodology with regular progress, planning and retrospective meetings to provide focus and set priorities in an agile way.
To deliver the best results and offer the right expertise at the right time, cross-industry insights from ML6 were leveraged and ML engineers from outside the project were involved in a.o. brainstormings, architectural design sessions & tech bytes on the latest ML development in the market. This led to new insights and contributed strongly to successfully building a state-of-the-art platform.