Aker Biomarine innovates with krill (and AI)

From the bottom of the food chain to the height of innovation

A tiny crustacean that’s survived in the world’s oceans for 130 million years may be the key to health for a lot of people — its oil is full of Omega-3 fatty acids. In the Antarctic alone, there’s enough krill to feed the ocean’s predators with plenty left over for human consumption.

The biotech innovator Aker Biomarine has turned this sustainable surplus into a science-backed line of nutrition products. “The krill advantage” isn’t just a brand tagline; it’s the key to business success, and Aker Biomarine has roughly 70% of the total global market when it comes to krill-derived nutrition products. In 2020, 1.6 billion doses of their Superba Krill were sold worldwide.

This Oslo-based life sciences company originally chose Box for secure, compliant remote data sharing. It now has upwards of 38 TB of data in Box and has discovered 119 threat alerts to date thanks to Box Shield. Building on Box as a content management platform, the company is now taking full advantage of Intelligent Content Management. Tord Isdal, Head of Digital Improvements, says “As a research and innovation organization, to be able to roll out AI, we have to be sure our data doesn’t go anywhere.”

Experimenting with AI across the business

“We’ve rolled out AI to everyone in the company,” says Isdal, “because we think it’s really important to play around with AI to understand how it can drive efficiencies and do a lot of other things.” Such experiments have resulted in varied uses of AI. For instance:

• Generating high-quality marketing content quickly and on the fly
• Translation of documents and meeting notes for teams in international offices, from Japan to South America
• Turning complex scientific research papers into lay language the marketing team can use to create content for podcasts and other types of assets

Throughout their testing of Box AI, active users have interacted with multiple endpoints across Notes, Hubs, Preview, Mobile, and Multidoc. Those users have averaged upwards of 10 prompts per week, with 70%+ favorability.

Aker Biomarine also takes substantial advantage of Box Hubs and its integration with Box AI, helping the IT and HR teams, for instance, more easily share and surface internal documentation. The R&D team also uses Box Hubs to aggregate files shared by clinical study partners in order to quickly extract insights across studies. This ultimately streamlines the product development process and helps uncover areas of opportunity to expand the company’s market presence.

Box Hubs helps decrease the time employees spend looking for information or duplicating content that already exists, and they can now find what they need even more easily in Hubs with Box AI. But the real ROI of AI will come from agents.

Agentic AI on the manufacturing floor

Aker Biomarine is now turning to agentic AI to enable more sophisticated workflows. As Isdal says, “As a manufacturing company, one of the most important things to us is uptime in our production facility.” AI agents built with Box AI can perform entire services to ultimately ensure continuity on the factory floor. For instance, agents can help reduce downtime by leveraging various sources of information:

• Handwritten reports and notes which contain valuable knowledge but are entirely unstructured
• A vast knowledge base of manufacturing-related articles from a vast knowledge base
• Live signals from IoT sensors (devices that collect data from machines and transmit it over the internet) in the factory

For the first time, the company can leverage both unstructured data and live signals within one system. By combining all of these sources of historic and live data, then configuring AI agents to extract the most important information and provide recommendations, Aker Biomarine can respond much more quickly to manufacturing issues. Enabling AI agents to talk to each other across maintenance systems and combine historical data (like the knowledge base) with live data (from IoT signals) — then write information straight to the data warehouse or to Box — gives Aker Biomarine the ability to solve problems in real time.

As Isdal describes, “Configuring an agent that can look at the information we have, then extract what the live signals are telling us and make decisions, creates an ecosystem around our production to increase uptime.”

The AI payoff in life sciences and manufacturing

Agentic AI opens up a lot of other potential future use cases for Aker Biomarine as well. Tasks that previously required a lot of manual coding, like predictive maintenance and production forecasting, could be handled more directly and efficiently by AI agents. This creates new possibilities for experimenting with the production process, including fine-tuning machine calibrations.

Isdal is particularly excited about the potential of agent-to-agent communication. This might look like custom agents that can write to Aker Biomarine’s cloud warehouse (GCP), then communicate to various services and apps within the rest of the technology ecosystem, ultimately writing information back to Box, where it can be securely stored and shared.

“This ecosystem around our production can save us a lot of money,” says Isdal. “But it's also going to be hugely valuable for our employees by enabling them to do the things they actually want to do —  work on the real problems. That’s where we’re seeing a lot of value.”
 

 

“We chose Box because of fit and security. And we've seen an increased relevance now with AI, where it's really important to know what happens with our data.”

— Tord Isdal, Head of Digital Improvements, Aker Biomarine