Ambitions are high when it comes to setting future data standards. They must be if the market is to keep pace with the changing technology landscape in adjacent sectors and deal with changing economic dynamics. These shifts include the increased rate of Environmental, Social and Governance (ESG) reporting or the complexities of efficient portfolio management in a geopolitically fragile world.
Ingesting datasets for improved risk modelling
The main limitation of many models on the market is that they work from the assumption that the future will be broadly the same as the past. Of course, this is untrue, whether a global pandemic, regional conflicts, or macroeconomic forces that cause high inflation rates.
For this reason, marine risk models need to be developed further to adjust to these changing landscapes automatically. This will be achieved by assimilating more complex and far-reaching datasets such as inflationary statistics and predictions, steel prices, shipyard utilisation and labour costs. Onboarding these macro-level insights will ensure the world of shipping and cargo is described most broadly and straightforwardly as possible, from seismic market changes to subtle operational changes.
Combining these datasets with the latest machine learning modelling techniques will allow a systematic and thorough approach to understanding the risk. This is made possible by accurately looking into the future to ensure sustainable capital placement and profitability.
An automated underwriting narrative
As we ingest these new datasets and create models that can describe risk from a macro and micro perspective, it becomes more and more complex to explain the modelled insights in a way a human can understand. Achieving this is critical to ensure the value of data is maximised and injected into what is still a relationship-driven market.
Because of this, advances must be made in further bridging this gap between the latest machine learning and artificial intelligence technologies and the existing market. In the case of our Quest Marine platform, for example, investments are being made into converting what was traditionally a ‘black box’ model approach into an events-based system that highlights fundamental changes to an asset or portfolio that are tangible to the end user. Combined with the machine learning modelling approach, an automated underwriting narrative will support the user in deploying these insights to impact the bottom line more effectively.
These events include changes in ownership, management structures or port visits. More complex scenarios include increasing regional shipyard utilisation rates and a new charter agreement with a different trading area, meaning that access to drydocking services has become more costly. An events-based approach to modelling will ultimately provide more detailed descriptions of the past and more accurate and dynamic future predictions – we see both as critical to supporting the market in maximising value from data.
This approach and advanced exposure management tools allow the market to respond to and manage significant events more efficiently. Only then can we fully embrace the benefits of digitalisation, vastly improving risk assessment capabilities and reaching complete confidence levels in how business is placed.
