iTromsø and Polaris Media were among the first AI pilot projects in Norway to receive significant media attention when the generative AI wave began sweeping across the world in 2023.  Together with the Bergen-based IT company Visito and IBM’s Client Engineering team, iTromsø developed an application - Djinn- that downloads and processes large volumes of documents daily from building permit departments in multiple Norwegian municipalities—extracting valuable news stories from the data.

Now, nearly three years later, the solution and the editorial data platform that spurred out from it, continues to deliver strong results for the media house and its journalists. It still attracts considerable attention within the media industry, despite the rapid pace of technological development.

Today Djinn is used by 36 local newsroom and the results are considerable:

·       From 2 hours to 15 minutes: Time spent reviewing building permit cases has been reduced by 80%.

·       Improved accuracy: Fewer cases are overlooked, and more are identified at an early stage.

·       Lower barrier to entry: New journalists can easily use the tool without prior experience.

·       Competitive advantage: Newspapers can be first to publish local news.

Key Success Factors to succeeding with AI implementation at scale, as identified by iTromsø:

·       Culture: People and processes are at the core.

·       Trustworthy AI Transformation: Fundamentally a change management process.

·       Early Involvement: Inclusion of subject matter experts—journalists in this case.

·       User-Centric Development: Designed with editorial needs in mind.

Governance:

·       Responsible use of data for journalism

·       Privacy by design and default

·       Data security and traceability

·       Bias mitigation

Collaboration and Recognition

Djinn was the result of close collaboration between IBM Client Engineering, Visito, and iTromsø/Polaris Media.

“Djinn and the platform developments that follow at Polaris Media are excellent examples of how AI can enhance and scale user experience,” says Tue Marker, Head of IBM Client Engineering for Denmark and Norway.

Internationally, iTromsø and Polaris Media are shaping how AI is applied in journalism. Their methods and results have been recognized and featured by leading journalism organizations and industry bodies, including WAN‑IFRA (World Association of News Publishers), INMA (International News Media Association), the Online News Association (ONA), and Nieman Journalism Lab. Their work has also been covered by major international media, including Der Spiegel, and cited in international conferences and peer‑reviewed academic research.

The iTromsø-team has also received multiple awards for their efforts, and the work and results has been published by various journalistic and research publications, such as Springer. Visito was also awarded with an IBM Partner plus Global Awards for Digital Labor in the EMEA Region, for their work with Djin.

IBM products used for Djinn:

·       PostgresSQL

·       watson.ai Studio

·       watson.ai Runtime

·       Code Engine

*Links:

·       Benton, J. (2025, November 4). A small Norwegian newsroom punches above its weight with a data‑driven, human‑centered AI strategy. Nieman Journalism Lab. https://www.niemanlab.org/reading/a-small-norwegian-newsroom-punches-above-its-weight-with-a-data-driven-human-centered-ai-strategy/

·       Elo Dean, S., Giske, L. A., Mostein, H. J., Podestà, S., Helland Barndon, H., Stegane, S., & Nordberg, H. (2024). Djinn—Data journalism interface for newsgathering and notifications. In Artificial Intelligence XLI (Lecture Notes in Computer Science, Vol. 15447). Springer. https://doi.org/10.1007/978-3-031-77918-3_11

·       Henriksson, T. (2025, November 4). A small Norwegian newsroom punches above its weight with a data‑driven, human‑centred AI strategy. WAN‑IFRA. https://wan-ifra.org/2025/11/a-small-norwegian-newsroom-punches-above-its-weight-with-a-data-driven-human-centred-ai-strategy/

·       iTromsø. (n.d.). iTromsø høster internasjonal oppmerksomhet [News article]. https://www.itromso.no/nyheter/n/pPJVaE/itromsoe-hoester-internasjonal-oppmerksomhet

·       IBM. (2024). iTromsø of Polaris Media: Client testimonial. https://www.ibm.com/case-studies/itromso-of-polaris-media-client-quote

·       International News Media Association. (2024). Djinn—Data journalism interface for newsgathering and notifications (Best Practice case study). https://www.inma.org/best-practice/Artificial-Intelligence/2024-409/Djinn-Data-Journalism-Interface-for-Newsgathering-and-Notifications

·       Langer, U. (2025, September 16). How iTromsø built Norway’s most AI‑ready newsroom. NewsMachines. https://newsmachines.beehiiv.com/p/how-itroms-built-norway-s-most-ai-ready-newsroom

·       Online News Association. (2024, August 20). Case study: Djinn, an AI‑powered data journalism interface. https://www.journalists.org/news/case-study-djinn-an-ai-powered-data-journalism-interface

·       WAN‑IFRA. (2025, October 30). “Masters of our own fate”: Media leaders embrace an AI future on their own terms. Public News. https://www.publicnow.com/view/E54BF5490FDE5B81D95041E6BB5035D33CE8EC19