Aquaculture has long sought ways to improve fish welfare while optimizing operational efficiency. Yet, with millions of data points generated across thousands of farm sites, the challenge has always been making sense of it all. At Manolin, we've built a data intelligence platform that harnesses 30 years of industry data - spanning over 30,000 cages - to forecast fish welfare using AI.
Recently, we presented this work at the Fish Vet Society Conference, demonstrating how structured data, predictive models, and AI-driven insights can reveal the underlying factors that make or break fish health outcomes.
The power of structured knowledge has shaped human progress for centuries. The Encyclopedia Britannica, first published in Edinburgh in the 1760s, aimed to systematize human knowledge, offering a reliable reference point across disciplines. Today, AI is revolutionizing this same idea—but instead of books, we are organizing aquaculture data into a system that enables farms to make better decisions.
Aquaculture data is inherently complex, spanning disease information, treatment types, and environmental conditions. By structuring this data, we can move from fragmented, anecdotal decision-making to a systemized approach where AI can detect patterns, predict outcomes, and optimize farm management.
Data in aquaculture is often categorized into specific variables, for example:
Disease Information – Pathogen type, duration, severity, environmental conditions.
Treatment Data – Treatment type, operator, temperature, duration.
Operational Variables – Boat traffic, production cycles, farm location.
With a systemized approach, this type of data becomes a foundation for AI models that can analyze trends and generate highly accurate forecasts, essentially creating an encyclopedia for aquaculture intelligence.
One key question is whether structured data improves predictive capabilities. To test this, let's look at an example where we analyze fish growth using three different data structures:
Each dataset was used to train a predictive model, which was then tested on unseen data. The results were clear:
Simply by improving the structure of the dataset - without altering model architecture - we increased predictive accuracy by 13%. This underscores the fact that clean, organized data is the key to meaningful AI applications in aquaculture.
While forecasting growth is a critical component of fish farming, AI can also be used to optimize treatments and reduce mortality events. Mechanical delousing treatments, such as Thermolicers, Optilicers, and Hydrolicers, have become industry standards - but not all treatments always yield successful outcomes.
The above graph shows over 5,000+ treatments mapped by Manolin. These are cleaned and grouped together to create treatments with similar fish profiles, using mortality and growth patterns of the fish. Once our AI clusters everything together, we can then isolate bad treatments and compare those outcomes (represented in above graph by the clusters that are outlined).
By analyzing treatment data across thousands of farms, integrating fish health records, and applying AI-driven clustering analysis, we can identify key risk factors contributing to post-treatment mortality. AI allows us to group similar treatments and isolate the variables that contribute to poor outcomes. This data-driven approach provides farms with an evidence-based framework to optimize treatment protocols and improve fish welfare.
One of the most compelling applications of this AI-driven approach has been our collaboration with Hofseth and Cargill. In 2023, we worked together to optimize their treatment strategies using structured data and predictive modeling. The impact was significant:
By implementing a structured approach to treatment planning and forecasting, Hofseth and Cargill achieved substantial improvements across both fish health and financial performance.
The industry is at a tipping point where data-driven decision-making is no longer optional—it’s essential. Moving forward, farms that invest in systemized data collection and AI-driven forecasting will have a competitive edge, allowing them to:
The ability to predict fish welfare outcomes with high accuracy represents a transformative shift in aquaculture. With AI, we’re not just reacting to challenges - we’re anticipating them, ensuring healthier fish populations and more efficient farming practices. Manolin is committed to leading this transformation, providing farms with the intelligence they need to optimize welfare and performance in every production cycle.
By leveraging AI and structured data, Manolin is transforming how farms predict, prevent, and optimize fish welfare - bringing aquaculture into a new era of actionable precision farming.