Ben Rose, Co-Founder and President of Supercede, the independent reinsurance platform, has warned that the reinsurance industry is not yet ready for a successful AI-driven future. According to him, the critical requirement for such success is the availability of clean, usable data.
In a recent interview with Reinsurance News, Rose likened reinsurance data to water. “If we think of reinsurance data as water, most of it today is undrinkable,” he said. “Any attempt to feed this unrefined data into downstream processes is likely to fail. By the time you reach the final product, you often discover that the data is simply unusable.”
Despite these challenges, Rose highlighted encouraging trends in data sharing among Supercede clients. Market participants are increasingly collaborating, creating the potential for what he calls ‘clean drinking data’, though he stressed that it is still a work in progress.
Rose said, “The adoption of shared infrastructure, combined with audit and validation tools, gives us a real chance at producing clean, reliable data across the industry. This could be a transformative step and may even make some AI-based initiatives feasible in the near future.”
Supercede has observed that large clients who previously used its tools primarily for individual data cleaning and tracking are now collaborating in real time. For example, cedents and brokers are jointly managing submissions, analysing limits, and exploring alternative approaches before going to market.
| Use Case | Description |
|---|---|
| Submission Data Cleaning | Ensures all client data is validated and structured properly |
| Deal Management | Tracks deals throughout placement, facilitating collaboration between cedents and brokers |
| Post-Placement Integration | Data flows into back-office systems for technical and financial accounting |
Rose also emphasised interoperability. Supercede has integrated with OpenTWINS, the primary back-office system used by many reinsurance brokers. This allows brokers to manage placements and post-placement data more effectively. Additionally, the platform feeds data into insurer back-office systems, enabling faster, more reliable access to critical information across the value chain.
However, he highlighted that poor-quality source data remains the largest barrier to adopting AI. “Training data for reinsurance AI is currently inadequate and may worsen outcomes rather than improve them. Practice makes perfect, but if you are practising on flawed data, it’s a serious risk.” He cited a recent MIT study showing that 95% of AI projects fail to deliver measurable returns, and warned that the reinsurance industry risks following this trend if data quality is not addressed.
The fragmented technology environment is another significant challenge. Many firms have developed in-house systems that do not communicate effectively, creating “technological islands.” Rose believes shared solutions between parties are essential to overcome this fragmentation, reduce friction, and allow for faster decision-making.
Supercede positions itself as the gold standard for data reliability in reinsurance. The platform ensures data is validated, audited, and maintained according to ISO 27001 standards, recently recertified for the seventh consecutive year. These security and compliance measures are critical to building the strong data foundations necessary for AI adoption.