This panel discussion from Data Jam 2024 covers creating a data ecosystem.
This panel discussion on creating a data ecosystem provided a thought-provoking exploration of how insurance organisations are wrestling with data fragmentation, AI adoption, and the evolving role of ecosystem thinking in driving operational and strategic value. It offered a balanced combination of practical experience, cautionary insight, and tactical guidance for anyone looking to transition from disconnected data assets to orchestrated value creation.
Defining the data ecosystem
The discussion opened by challenging the popularised use of the word ‘ecosystem’ as a fashionable substitute for partnership networks. While the term is often misused, the panel clarified that a true data ecosystem involves orchestrating a dynamic interplay of systems, data sources, platforms and people, all aligned toward customer value. In the insurance context, this includes integrating underwriting, claims, and customer operations through seamless data flows that enable intelligent, adaptive decision-making.
Participants reflected on the integration challenge, where the proliferation of technologies, vendors, and platforms creates hidden complexity. Legacy architectures and siloed systems add further friction. Even well-funded transformations can stall when they fail to address foundational data architecture. One panellist, referencing experience in a major personal lines insurer, highlighted how transformation at scale only succeeded when underpinned by a robust data model capable of feeding operational and customer-facing systems in real time. Orchestration, not simply aggregation, was the cornerstone of success.
AI as an enabler, not a shortcut
AI was positioned as an enabler, not a panacea. While large-scale automation, faster data discovery and entity resolution are increasingly possible, the panel stressed that AI is only as useful as the underlying data foundations. Poor data quality or disconnected records cannot be solved by machine learning alone. Participants shared their experience of deploying AI-enhanced tools to infer relationships, accelerate metadata tagging and identify joins between previously isolated data sources. These advances help accelerate discovery, but require validation and human oversight to ensure robustness and reliability.
Several speakers underscored the enduring value of foundational data initiatives. Master data management, data quality improvement and the creation of internal data standards were all described as essential precursors to successful AI deployment. There was an acknowledgement that AI tools are accelerating data structuring, but they must be aligned to a broader strategic framework or they risk becoming point solutions without scale. The analogy of "putting the cart before the horse" captured the temptation to chase AI capabilities without first ensuring consistent, trustworthy data across the organisation.
Strategies for implementing a data ecosystem
In discussing practical steps, panellists outlined strategies for initiating data ecosystem development without falling into the trap of over-engineered, slow-moving programmes. One recurring theme was the importance of experimentation. Many noted that proof-of-concept pilots, delivered quickly with a defined scope, help build momentum and reduce resistance. Iterative delivery also enables boards to tolerate early missteps if the business case is clear and improvements can be demonstrated rapidly. A fail-fast approach, framed around marginal gains, was seen as key to sustaining executive sponsorship.
The discussion turned to the cultural conditions necessary to support this experimentation. Psychological safety, clarity of vision, and alignment with business strategy were described as critical. One speaker noted that the board only truly listens when the proposition is clearly articulated in terms of increasing revenue, reducing cost, saving time or mitigating risk. The ability to tell a compelling story that aligns data and AI initiatives with strategic goals is crucial in winning buy-in and sponsorship.
Creating innovation through intrapreneurship
The theme of intrapreneurship recurred throughout the session. The panellists stressed the importance of cultivating internal entrepreneurs who are willing to experiment, take calculated risks and push against organisational inertia. However, scars from past failures have often led to risk aversion, particularly in the London market. Large-scale investments in technologies like blockchain or unproven InsurTech platforms have, in some cases, created caution around bold initiatives. As a result, boards are more willing to support incremental change than transformational bets.
Yet, the panel also recognised the danger of being too incremental. The concept of ‘marginal gains’ has appeal, but there was a challenge from the audience about whether the insurance industry risks losing sight of the 40 percent opportunity while focusing only on two or three percent improvements. While some felt large-scale gains could be the product of accumulated marginal improvements, others argued that ambitious bets are necessary to unlock significant competitive advantage. The risk, however, must be proportionate and aligned to a tested business case.
Delivering value with discipline
To navigate this balance, the panellists proposed a model of delivering early wins while painting a plausible, strategically aligned vision of long-term value. Mapping customer journeys, identifying specific friction points and deploying targeted AI solutions against those pain points allow organisations to demonstrate progress while building toward a broader goal. This disciplined focus enables boards to release incremental investment and fosters belief in a longer-term roadmap.
Conclusion
Finally, the conversation explored how to overcome internal resistance. One participant described the importance of understanding the root causes of objection, particularly from technical or compliance stakeholders. Building relationships across the business and listening to concerns about data privacy, security or regulatory compliance ensures that innovation does not become a proxy for recklessness. Another noted that change only becomes viable when initiatives are framed in language the business understands, not in technical or AI-specific terminology.
In summary, the session painted a detailed and realistic picture of what it takes to build a data ecosystem in insurance. It requires more than technology partnerships or AI experimentation. It demands strategic clarity, strong data foundations, cultural openness and a pragmatic operating model that aligns with business outcomes. AI has a role to play, but it is not magic. Real value will come from combining experimentation with focus, ambition with realism and data quality with technological acceleration. The ecosystem is not a destination. It is a continuous, adaptive process of connecting what matters to deliver what counts.
Sign Up to TINsights
Where we share our latest blogs, industry reports and insights
