The insurance transformation agenda: Perspective from a roundtable on data, AI, governance and growth

How do you scale AI when your data foundations and technology stack are still being rebuilt or replatformed? That was the question on Wednesday 14 October, when Robin Muir and I hosted senior insurance leaders at our London office: transformation executives, chief data officers and technology specialists.  

The private roundtable, ‘The Insurance Transformation Agenda: Data, AI, Governance & Growth’, was run jointly by ECMS managed solutions and the specialist contingent recruitment arm of Eames Group, Eames Consulting and covered four themes: 

• Replatforming strategies and lessons learned from the process
• The shift from data and analytics to AI-first approaches
• Maturing governance and operating models in the era of accelerated automation
• Balancing AI at scale with foundational data investment 

Discover key takeaways:

Replatforming strategies in insurance and lessons learned 

Older insurance organisations originally built their platforms for data analytics, not AI. This has driven substantial investment in migrating data, applications and systems, as organisations seek to transform legacy data estates and achieve targeted optimisations. 

However, as one leader put it, replatforming alone does not deliver business value. When a problem is treated as a platform problem by default, the true business value may be skipped.  

Starting without clearly defined business outcomes, what the business wants to achieve, may result in a costly data lakehouse, an expensive digital store that doesn’t in itself ensure data interoperability, so that CEOs wonder what return they gained from platform spend. 

From data and analytics businesses to AI-enabled organisations 

Although insurance is good at using the language of agility, data and AI leaders say it is not always as good at putting it into practice. On the other hand, for insurers who are more advanced in their automation journeys, there is also a question mark over how embedding AI now will impact and price into future cost models. 

As I wrote in the September 2026 issue of the ECMS60 newsletter, future costs aren’t always priced in when insurers implement AI. Leaders suggested that remains true.

Organisations are still working out the future cost model of AI: how scalable it is, and how locked in they will become to a particular model or ecosystem.  

The pressure to move anyway is considerable. One leader observed that fear of being out of date is itself becoming a blocker to progress. They added that the industry often talks about AI as the solution rather than part of it and discusses it separately from business objectives. 

What the most successful firms have in common, in ECMS’ delivery experience, is that they hunt and solve defined problems, rather than simply sprinkling automation on top of systems that don’t fully support it.

People and capability: How technology-led transformation delivery squads have changed 

Although not on the official agenda, people and capability questions inevitably came up. Data and AI leaders shared that resistance to AI adoption still often lies more at the senior end, while junior workers entering the market are increasingly AI-native.

A key takeaway was that Proof of Concept (POC) teams in insurance are now structured very differently to how they were a few years ago, with squads often consisting of very young AI-native technology specialists plus an experienced insurance business analyst, project manager, and subject matter expert.  

How data and AI governance and operating models are evolving as automation accelerates 

Previously, the conversation focused on uneven data quality and interoperability. Now, governance is increasingly top-of-mind. Data and AI leaders confirmed that governance maturity is still uneven across organisations, as we found at a previous data roundtable held with 12 CDOs this year. 

Some firms do not yet have sufficient governance for the degree to which they have embedded automation which creates audit risk. Firms are positioning the need for governance as regulatory protection, rather than red tape that would curtail ongoing use case exploration and process optimisation.   

Leaders in attendance argued for governance by design (rather than by email), with the same identity access controls applying to people and AI. With IT often having taken over governance, leaders are concerned about gaps between policy and practice.  

Insurance has often lagged behind other financial services in identity access management with persistent issues. 

In summary

These events bring data, AI and insurance leaders together to compare notes candidly on what is delivering lasting value and where the industry still needs to evolve. They provide valuable opportunities to share the kind of peer intelligence that fosters a stronger market for all. Thank you to everyone who joined us and shared their perspective so openly. 

Our full report on the insurance agenda concerning data, AI, governance and growth is coming soon. Combining insights from roundtables across 2026 with our latest research, it will explore how firms are scaling AI while protecting foundational investment, and where data and AI leaders plan to focus spend next.

---

If you would like to receive ECMS’ original research or an invitation to events for data and AI leaders, get in touch at contact@ecmanagedsolutions.com.