Tysers, a leading, specialist insurance broker, reports that the insurance industry, has seen a significant shift towards automation and data-driven decision-making.

Algorithmic underwriting is at the forefront of this transformation - a process that uses advanced algorithms, machine learning, and artificial intelligence (AI) to assess and underwrite insurance risks.

Algorithmic underwriters leverage vast amounts of data from various sources, including historical claims data, real-time market information, and external data. By processing this data through sophisticated models, algorithmic underwriters can identify patterns and correlations that may not be immediately apparent to human underwriters, providing at times, a cost-effective approach.

However, the rise of these platforms brings concerns. Tom Walker, Director of Hull and War in the Marine team at Tysers Insurance Brokers, explains: “Ensuring the quality of data is paramount, as inaccuracies can lead to flawed models and incorrect risk assessments. For example, Lloyd’s, a global insurance market that underwrites complex and specialised insurance risks is highly regulated, and the transparency of algorithmic decisions is crucial for maintaining trust with customers and regulators. Furthermore, through the implementation of algorithmic underwriting, nuanced knowledge and years of specialist expertise gained from human underwriters can be lost in translation. The industry must not let this happen as risk is not getting simpler.”

Tom adds: “In the shipping market the introduction of more modern, technologically advanced ships has heightened the need for seasoned shipping expertise within the underwriting community and illustrates that the presence of strong, knowledgeable and supportive insurers will be even more crucial in addressing such sophisticated risks. It’s important to remember that while AI enables efficiency and innovation, it is not a standalone model. The real value is driven by the intelligent combination of AI models and human expertise.”

More like this…

View all