AI x Insurance & Risk Management [Meet-up in Paris]

On April 4th in Paris, astoryaVC and BGV co-hosted a meet-up focused on Artificial Intelligence in insurance and risk management. Attended by approximately 50 individuals from insurance & banking corporates, as well as VC investors, the event provided a platform to delve into the tangible benefits these technologies offer across the insurance value chain. To provide practical insights, we were privileged to host presentations from several startups, offering firsthand accounts of how they harness data and algorithms to drive value within these sectors. It helps us address a pivotal question: “How can insurers get the most of AI technologies to master operational effectiveness and tackle emerging risks?”

Introduction: the Intersection of AI and Insurance

As a reminder BGV is a Venture Capital firm headquartered in Silicon Valley with offices in France, Israel and India. Specialized in cross-border investments, BGV invests in early-stage AI-based enterprise softwares shaping the future of the autonomous enterprise. astoryaVC is investing in tech-first startups which aim at revamping insurance and tackling emerging risks. It focuses on European companies with an available product (i.e. Seed stage).

The event kicked off by highlighting the explosive growth in GenAI investments in 2023, which saw a fivefold increase compared to the previous year. A significant portion of this funding—approximately two-thirds—was directed towards companies specializing in Large Language Models (LLMs), backed by major corporations like Google, Microsoft, and Amazon. This trend is largely attributed to the capital-intensive nature of developing LLMs. Furthermore, the industry is witnessing a surge in applications being built on these foundational models, reminiscent of the app development boom following the advent of smartphones. BGV and Astorya are focusing on the application layer, where specialized language models (SLMs) are being developed to leverage LLMs. These SLMs incorporate proprietary and domain-specific datasets, alongside workflow automation, to create targeted solutions.

Discussions centered around potential applications that can be built atop these AI models in the insurance industry:?

• Firstly, it promises to enhance operational efficiency across various sectors, with insurance ranking consistently among the top three industries expected to benefit from AI implementation. This alignment is logical, given that many incumbents prioritize operational streamlining as part of their short-term strategic objectives.
• Secondly, AI presents an opportunity to enhance insurers' capabilities in understanding, assessing, and pricing emerging risks, which have seen a notable uptick in recent times (e.g., climate change, cybersecurity, digital assets). Addressing these emerging risks entails the identification and utilization of new data sources, along with algorithmic analysis to unlock additional insurance capacity.?
• Lastly, the adoption of AI introduces a spectrum of new risks, spanning from issues related to intellectual property rights to potential algorithmic biases or system downtime. Nevertheless, these risks also signify opportunities for the introduction of innovative insurance products and the exploration of new market segments.

Keynote: “Lessons in AI: From POC to production” by Grid Dynamics

In the first item on the agenda, the event featured a compelling keynote by Mark Lister from Grid Dynamics (NASDAQ: GDYN), a BGV portfolio company, helping large enterprises and startups in their digital transformation journey by building digital platforms powered by data, cloud and AI. Mark provided invaluable insights based on his extensive experience in the field, focusing on the pivotal journey of transitioning AI projects from Proof of Concept (POC) to production—a critical hurdle for many in the industry.

The key takeaways:?

• Transitioning POCs to Production is rare but critical: Only 2% of AI POCs successfully make the leap to production.?
• Enterprises need to view AI deployments as part of a broader data strategy and this needs operational maturity to be successful
• FinOps is a key focus area when moving LLM tools from POC to Production. Techniques like model orchestration and token optimisation are necessary to keep costs as low as possible. Grid Dynamics shared a real life cost reduction roadmap where they succeeded to reduce LLM API costs from $3m to $1m annually keeping the same performance.

Startup pitch session

Discussions transitioned from macro insights to applications, with startups such as Zelros, Dattak, Greenly and WeatherMind taking the stage. Each presentation revealed a facet of AI's applicability—from enhancing customer interactions and streamlining claims processing to advancing cyber risk assessment and fostering sustainable business practices.

• Christophe Bourguignat , CEO and co-founder of Zelros , showcased how Zelros empowers insurance agents and brokers with data-driven decision making tools. By harnessing the capabilities of LLMs, Zelros offers real-time personalized insurance recommendations across channels. With more than 50 000 agents using Zelros’s Insurance Copilot today, the ROI is real, evidenced by one customer saving 10,000 hours monthly and another achieving a conversion rate increase of over 60%.
• Charlotte Couallier , CEO and co-founder of Dattak , shared concrete examples of how hackers are using GenAI in their attacks. With less than 3% of SMBs insured, Dattak's innovative use of AI in cyber insurance showcases the industry's readiness to tackle modern threats. Dattak combines cybersecurity services with the most comprehensive insurance coverage available, all in one product.
• Alexis Normand , CEO and co-founder of Greenly , showcased the automation of carbon management through AI, making it accessible to companies of all sizes. Greenly's approach significantly reduces the cost and time associated with GHG reporting—achieving a tenfold improvement over conventional consulting methods. Beyond mere carbon accounting, Greenly offers its 5,000+ customers best practices and actionable strategies for sustainability.
• Mehdi Gaaied , CEO and co-founder of WeatherMind , explained their computer vision-based solution designed to assist insurers in evaluating property damage following natural disasters. Initially focusing on drought risks, the technology allows insurance partners to pre-assess damages, ensuring that the scope of work is matched to the complexity of the case. This innovation significantly accelerates claim processing, shortening claim cycles by an impressive 30-fold!

Conclusion

These examples highlight various applications of artificial intelligence in insurance and risk management, showcasing tangible, real-time outcomes. It was crucial for us to concentrate not just on the potential future capabilities of technology but also on the concrete benefits it currently provides to established firms collaborating with external innovators. The results shared at this meetup serve as valuable data points for the internal debates often encountered by incumbents, particularly regarding the 'make or buy' dilemma. Additionally, observing how peers gain advantages through partnerships with startups should offer insightful guidance.

In case you’d like to learn more about these players, feel free to reach out directly to them:

• Mark Lister from Grid Dynamics
• Christophe Bourguignat from Zelros
• Charlotte Couallier from Dattak
• Alexis Normand from Greenly
• Mehdi Gaaied from WeatherMind

Looking forward to meeting you at our next meetup !

Sarah Benhamou & Florian Graillot