Media Info Groep

Online media era requires faster search response

Rather than just provide a news clipping service, Media Info Groep gives its clients the story behind the news, providing insights into trends and noting differences in tone of voice and sentiment. It delivers real-time content through online portals and dashboards. As time went on, and online media began to grow in importance, the number of sources grew from some 2,000 to over 100,000, requiring a sophisticated content archive, hosted in a Microsoft SQL database. It was at this point that the company decided it needed a more robust solution to search its database of over 250 million documents.

Mark Reisz, CEO of Media Info Groep, took up the story: “The biggest challenge wasn’t even the number of documents, but the fact that every day we store up to 80,000 new documents. Whereas years ago, clients might have been happy to wait a couple of days for their news results, today a much more instant response is expected. This is where [OpenText Knowledge Discovery] came to the rescue for us.”

[OpenText Knowledge Discovery] provides us with a solid and reliable service that our entire business model depends on. Without [it], we could not deliver quality content to our clients in a timely manner.

Searches 250 million documents to deliver results in seconds

OpenText Knowledge Discovery provides a single processing layer that gives users a real-time contextual understanding of both structured and unstructured data. This enables Media Info Groep to access many different media sources to monitor and analyze the information in a single user interface. For radio and TV, Media Info Groep uses a speech recognition engine. This advanced audio recognition technology is part of OpenText Knowledge Discovery and enables users to understand and analyze audio content in several ways, from speech recognition to sentiment analysis. Speech is translated into text, and subtitles are also used to input data into the database. Media Info Groep uses tools to scrape social media platforms. The content is then converted into text format and ingested by OpenText Knowledge Discovery for further analysis.

“Leveraging [OpenText Knowledge Discovery], we process up to 35,000 search requests every 15 minutes”, said Reisz. “We have a total of over 30,000 different search profiles, which [OpenText Knowledge Discovery] executes. These range from simple searches to very complex ones with many variables, for instance, for our energy market clients. When human filtering and analysis is added to the process, results are consolidated and typically returned to the client within an hour. Simple search requests are fulfilled in just seconds. This compares to days and hours prior to our [OpenText Knowledge Discovery] adoption.”

Many Media Info Groep clients publish daily newsletters that summarize and consolidate their news coverage. They just want 10–15 of the most relevant articles to include in this. Previously, Media Info Groep’s reader team would manually process a far wider content selection and decide on the final content. Today, OpenText Knowledge Discovery completely automates this process through its AI capabilities to meet the increasingly sophisticated levels of service required by Media Info Groep clients and publish timely and relevant newsletters. “Our clients might like to know that they are regularly mentioned in conjunction with certain competitors,” explained Reisz. “Having related news items is useful information for competitive benchmarking and analysis.”

Whereas years ago, clients might have been happy to wait a couple of days for their news results, today a much more instant response is expected. This is where OpenText Knowledge Discovery came to the rescue for us.

Adding value with AI capabilities

OpenText Knowledge Discovery uses AI-driven sentiment analysis and named entity recognition, helping Media Info Groep clients understand how they are being presented to their customers, identify trends, and pinpoint regional sentiments. This enables them to quickly react to any negative information. “It still impresses me every day how we run an OpenText Knowledge Discovery query against a database of hundreds of millions of documents and receive the results within a couple of seconds,” commented Reisz. “And not just a simple result listing, but a helpful summary and relevance ranking. We reduced our reader team by half and still produce the same quality of content."

He concluded: “OpenText Knowledge Discovery provides us with a solid and reliable service that our entire business model depends on. Without [it], we could not deliver quality content to our clients in a timely manner.”