About
QBE’s Cyber Submissions AI Assistant is an innovative solution that leverages generative artificial intelligence (AI) to reimagine underwriting.
Innovation presentation
QBE’s Cyber Submissions AI Assistant is an innovative solution that leverages generative artificial intelligence (AI) to reimagine underwriting. By tapping into large language models (LLM) and generative AI capabilities, the initial risk assessment of cyber liability submissions is performed more than 60% faster. This significant reduction in the time and effort it takes to evaluate submissions enables underwriters to focus on more submissions and opens the door to increased GWP.
QBE’s Cyber Submissions AI Assistant launched last December and was developed to accelerate the pace of the broker submission review process while enabling Cyber Underwriters to select the best risks.
The Cyber Underwriting team receives a large volume of submissions each year and this creates a challenge in being able to review them all in a timely manner. With 40% of an underwriter's time spent on manual administrative tasks, including validating and processing submission data, coupled with 80% of data being unstructured and not machine-readable, the strategic opportunity to improve the process was clear.
Focusing on this opportunity to improve broker submission responsiveness and the quality of risk selection, the Cyber Submissions AI Assistant team was built. The team included representatives from the Underwriting business, Technology Services, Group Chief Data Office and QBE Ventures – a true demonstration of collaboration across the enterprise.
Working closely with the Underwriting business to uncover pain points, AI was leveraged to build the solution. Generative AI capabilities are used to package and automatically route emails and documents received from brokers, and again to extract relevant data. This is integrated with other data points to enable a comparative analysis of risks, providing a first-level assessment including citations and justifications for its findings. Machine learning is then leveraged to compute win propensity scores.
In the space of 12 weeks, the team moved through various stages to get the Cyber Submissions AI Assistant into production.
Uniqueness of the project
Evaluating policy submissions quicker was a game changer for QBE. While the industry is exploring how to leverage generative AI, QBE was able to leverage the technology for our submission process to generate efficiency and add to bottom line growth. Underwriters are already saving time (60% faster), quoting more, managing risk better and increasing the win ratio.