Member event: AI and automation online discussion

Summary 

AI Special Interest Group Meeting 26th February 2025

Topic: Advancing AI Adoption in the London Market – Key Next Steps

 

1. Meeting overview

The meeting was structured as an informal discussion focusing on two primary areas:

Market-wide challenges in AI adoption and data standardisation
Internal applications of AI for efficiency and business value
The session included a recap of a previous TINtech London Market AI discussion, key takeaways from that session and an open forum for participants to share insights and challenges.

2. Market-wide challenges in AI and data standardisation

2.1. Importance of data consistency

Participants highlighted that AI adoption relies on high-quality, structured and standardised data. The current lack of consistency in how data is collected and exchanged within the market presents significant barriers. It was noted that:

Brokers and carriers collect data independently, leading to mismatches.
Efforts to establish a shared data standard (e.g., Blueprint 2) face implementation delays.
Data must be auditable and explainable to meet regulatory requirements.
2.2. Broker reluctance to standardise data

There was discussion around brokers’ unwillingness to provide data in a standardised format, particularly before the placement process. Challenges include:

Brokers often prefer to retain flexibility in data submission.
They may see little direct benefit in providing structured data.
Backloading of data post-placement remains common.
2.3. The role of market initiatives and Blueprint 2

While some participants expressed optimism about the potential of market-wide data initiatives, concerns were raised about previous failures to implement such solutions effectively. It was noted that:

Market-wide initiatives require cooperation but have historically struggled due to competing priorities.
Some insurers are reluctant to invest in systems that could soon become outdated.
The need for incremental progress was emphasised rather than attempting to overhaul processes in a single step.
2.4. Potential market-wide solutions

Suggestions included:

A centralised data ingestion service offered by the market to eliminate duplication of effort.
Encouraging brokers to use structured data by demonstrating efficiency gains.
Exploring smaller, targeted partnerships between insurers and brokers to develop aligned solutions.
Despite these ideas, there was consensus that large-scale implementation remains challenging due to misaligned incentives and a lack of industry-wide cooperation.

3. Internal AI applications for efficiency and business value

3.1. AI in document ingestion and data extraction

Many organisations are focusing on AI solutions for automating document processing to improve underwriting workflows. Key points included:

AI is being used to extract and structure unstructured data from emails, PDFs and broker proposal forms.
Automating data extraction allows underwriters to focus on decision-making rather than manual data entry.
Some organisations are considering whether to build in-house solutions or rely on third-party providers.
3.2. AI for underwriting efficiency and risk selection

Some participants highlighted that their focus is on using AI to improve risk selection rather than administrative efficiency. Use cases discussed included:

Leveraging AI to analyse external data sources for better underwriting decisions.
Using machine learning to refine portfolio management and risk assessment.
Exploring AI-driven analytics for real-time risk monitoring.
3.3. AI in claims processing

AI was identified as having strong potential in claims management, though most participants indicated that their current focus is on underwriting applications.

3.4. AI in customer service and decision support

AI-driven chatbots and virtual assistants were discussed as potential tools for:

Supporting junior brokers by providing real-time underwriting guidance.
Enhancing customer interactions through automated responses and decision support tools.
Providing enriched reporting with AI-generated insights and context.
 

3.5. Challenges in AI adoption

While AI offers clear benefits, challenges include:

High implementation costs, particularly for smaller or lower-volume businesses.
The need for quality training data and ongoing model refinement.
Ensuring AI-driven insights are trusted and explainable to stakeholders.

4. Next steps and upcoming events

A Blueprint 2 Special Interest Group is available for those interested in further discussions on data standardisation and implementation challenges.
Breakfast briefings will be scheduled to continue conversations on AI use cases and best practices.
Upcoming industry events include a session on AI in insurance on 11 March and additional discussions leading up to the TINtech conference in June.
Participants were encouraged to:

Share case studies and examples of successful AI implementations.
Suggest topics for future meetings, which may focus on specific AI applications in underwriting, claims or operations.
Engage with market-wide initiatives while continuing individual AI adoption efforts.
The meeting concluded with a commitment to ongoing collaboration and knowledge-sharing to drive progress in AI adoption and data transformation in the insurance sector.