Sentina
Sentina is a long-standing Finnish content provider for elderly care services, offering daily media programs that supports memory, wellbeing, and engagement. Their platform, Sävelsirkku, which has run successfully on AWS for years, delivers a wide range of audio and visual content, from music and historical moments to exercise routines and nature videos.
Recently, Sentina saw a growing need to make that content easier to search, tag, and personalize — both for staff and their increasingly digitally-savvy residents. So, they called in NordHero to help.
“In daily care work, quickly finding relevant, engaging content makes a huge difference in saving resources for interaction with residents,” says Kari Mikkonen, Sentina’s CEO. “And we were looking for more precision and efficiency.”
Using AWS and generative AI tools, NordHero set up a project to start upgrading Sentina’s media systems. Sävelsirkku has a vast content archive for their clients to enjoy, for example, music and exercise routines to historical news and radio clips. Things like a radio broadcast from the 1952 Helsinki Olympics, describing the moment when Paavo Nurmi lit the Olympic flame, or a favorite childhood song performed by a well-known singer. For residents, such input can trigger vivid memories of sitting with the family around the radio or joining the roar of an Olympic crowd, bringing the past back to life.
However, Sentina’s content was created by different people over 25 years, with uneven sound quality and almost no searchable metadata. This meant that finding the right piece of media for a specific date or occasion was often slow, manual, and of varying quality.
Sentina and NordHero started with the most urgent challenge: the audio archive.
“The content itself was incredibly valuable, but there was a huge variation in quality and structure,” recalls Pekka Malmirae, CEO and cloud solutions architect at NordHero. “Our goal was to make it more consistent and searchable without disrupting the way Sentina already worked with their elderly care customers.”
The first step was to use AWS media services to normalize the sound by making sure there was consistent volume and quality across all recordings. Next, NordHero used AWS Generative AI tools to automatically transcribe the content, summarize it, and extract keywords for search functions.
Once the automation pipeline was in place, the team could process more than a thousand audio files quickly. “It didn’t require a massive rebuild, just the right tools in the right order,” explains Arttu Pekkarinen, NordHero’s Cloud Architect, who led the team on the technology side. Then with the foundation laid, the system could start to run itself with minimum manual effort. “We were surprised by how easily Generative AI capabilities could be integrated and automated into our existing systems,” Mikkonen says.
This has also laid the groundwork for further upgrades. For example, the transcriptions can ultimately be used as subtitles for the hearing-impaired, automated translations for international users, or content tagging for personalization. This development enables, for example, the continued enhancement of search functionality, which has already been initiated.
While the Sentina project was going on, Malmirae points out that the BBC World Service was making headlines for a similar project, but on a much larger scale. The global broadcaster migrated its century-old media archive to AWS in a massive digitization project that also aimed to improve the quality and accessibility for its one hundred years of video and audio production.
“Even though Sentina’s case is not the same as a big broadcaster like the BBC,” Malmirae says, “it was interesting that they were using many of the same AWS tools for their content upgrade, and with many similar goals.”
In this context, Malmirae praises the AWS platform for its democratic approach. “It’s a truly level playing field, which means we can bring the same advanced capabilities to clients of any size. You don’t need a big IT team or an endless budget, just the right use-case and a team that knows how to navigate the AWS ecosystem.” And that’s where NordHero comes in.
Sentina’s clients in Finland and Sweden are increasingly tech-savvy: many use smart devices and are comfortable with digital tools. Soon both residents and caregivers will be able to search Sentina’s content archives more efficiently themselves using dates or key words based on personal interests, making daily programming more tailored.
“The outcomes of this project will be used to further develop the search functionality,” says Mikkonen. “Finding suitable content quickly is extremely important in caregivers’ daily work. That’s why we’ve been developing a science-based daily activity model with professionals of the field for over 25 years. The model helps caregivers to create daily doses of joy only when they find contents efficiently and in the correct moment.”
In everyday care, staff often need to locate additional material on the fly, whether it is a specific rehabilitation need, a resident’s personal interest, or simply a change in mood. In that context, says Mikkonen, “the quality of information in the backend system is essential.”
Sentina is already fielding requests for more personalized and localized content. The company has also set its sights on scaling into new languages and markets. Thus, more work is to be done, and Malmirae and his team would be happy to continue the good collaboration.
“The collaboration has been smooth, professional, and solution-oriented,” Mikkonen affirms. “NordHero brought deep AWS expertise and the ability to quickly adapt to our needs. Communication was easy, and they took active responsibility for driving the project forward.”
Thinking about an in-house AI-driven project of your own? Click here to connect with NordHero team.
