We all know effective AI is dependent on clean, well-structured data. While data hygiene is essential, the real game-changer is enriching data with tacit knowledge. Tacit knowledge is the expertise and decision-making instincts of experienced professionals, which AI can’t replicate.

Why Data Hygiene Isn’t Enough

To keep AI-driven decisions relevant and effective, insurers must:

• Update data continuously to keep it relevant for changing risk landscapes.
• Enrich datasets by combining internal and external sources.
• Convert unstructured data into actionable insights using advanced AI tools.

These steps improve data quality and may boost model performance, but will never capture the deep expertise of experienced insurance professionals.

Turning Expert Insights into AI Power

Many decisions in P&C insurance rely on subjective human judgment, like identifying nuanced risks or anticipating unforeseen claims, and AI can’t learn this from historical data. However, insurers can embed expert reasoning into AI workflows by:

1. Flagging Unusual Cases: When AI suggests an answer, experts can review and refine it.
2. Recording Expert Decisions: AI can log and learn from these adjustments over time.
3. Build a contributory data model: Insights from experienced professionals become structured data for future decisions.

The Value of Institutional Memory

As long-time employees retire or leave, their knowledge often goes with them. By capturing it within AI systems, insurers can:

• Preserve expertise for future teams.
• Improve AI recommendations with real-world insights.
• Ensure consistency in decision-making.
• Support compliance and audits with clear, documented reasoning.

AI That Learns From Experience

Real success in AI will come from combining automation with human expertise. AI should not just process data but learn from the company’s best decision-makers. By embedding expert knowledge into AI systems, insurers can stay competitive, reduce risk, and make smarter decisions faster.

Real-World Applications

A recent report from Cake & Arrow outlines some real world examples of AI can capture tacit knowledge.

1. Policy Audits: AI can learn how seasoned brokers refine policy language and identify coverage gaps. It can then offer real-time guidance to new brokers during policy reviews, ensuring consistency and reducing errors.
2. Claims Adjustments: AI systems can help newer adjusters make faster, more informed decisions by analyzing how experienced adjusters evaluate claims.
3. Underwriting: For complex cases, AI can capture the strategies of senior underwriters and provide dynamic, context-aware recommendations, helping less experienced team members improve their decisions.

^1,2 Cake & Arrow (2024, November 4). A new golden age for insurance? How AI can help make insurance a more human experience. Retrieved January 24, 2024, from Cakeandarrow.com