The insurance industry is not short of ambition to innovate. Blockchain-enabled claims management. AI-powered personalisation. Real-time fraud detection. The appetite for innovation is real, and growing.
But wanting to innovate and being ready to innovate are two very things.
For years, the industry has been cautious about change. That caution is now harder to justify. The technologies on offer are too powerful, and the competitive pressure too great, to ignore. Yet a gap is widening between what insurers want to achieve and what their operations can actually support.
The infrastructure underneath simply was not built for this.
The ambition gap
The scale of that gap is striking. 82% of insurers believe AI will define the industry’s near future, yet only 14% have fully integrated it into their financial operations.
The barriers are clear. 42% of firms cite legacy system integration as a key challenge. 39% point to fragmented data environments. 40% lack the in-house AI expertise to move forward.
Beneath all of these sits the same root cause: data. Specifically, the absence of clean, well-governed data.
More than half of insurance firms describe their data governance frameworks as early-stage or still developing. The consequences are visible in operational budgets, where large portions are absorbed, correcting errors generated by manual processes. That is capital that could be funding innovation but instead is absorbed by operational inefficiencies.
In this kind of environment, deploying AI on top of existing systems is not a strategy, it is a risk. AI learns from the data it is fed. When that data is inconsistent, incomplete or ungoverned, AI does not just underperform, it amplifies the dysfunction already present in those systems, but at scale.
The 6% of firms reporting no AI usage at all may, in some cases, be making a more considered decision than those rushing to deploy without the foundations in place to support it.
Automation needs strong foundations
There is a temptation to treat automation as an add-on, something that can be dropped onto existing workflows to smooth them out. In a fragmented data environment, that logic fails quickly. Automation does not resolve underlying complexity. It embeds it.
Blockchain’s potential in insurance is real: faster claims handling, stronger fraud prevention, instant proof of coverage, and improved access for underserved customers. All of this is enabled by a distributed ledger that creates a shared, tamper‑proof record of events. But that same immutability becomes a liability when the underlying data is flawed. Misaligned bordereaux records, incorrect claim references, or inconsistent policy data don’t just create friction, they become permanently embedded in the chain, making reconciliation and root‑cause analysis far more complex.
Ultimately, blockchain’s value depends entirely on the accuracy, alignment, and governance of the data flowing into it. The same logic applies to personalisation. AI and advanced analytics can help carriers identify customers at risk of attrition, create tailored offers, and anticipate claims. But only when trained on reliable, governed data.
Innovation does not fix a data problem. It exposes it. Firms that have not invested in clean reconciled data will find themselves struggle to innovate!
Building for what comes next
The firms seeing the strongest returns from innovation are not the ones moving fastest. They are the ones investing in the operational infrastructure that makes sustainable innovation possible.
That means automated reconciliation to ensure data accuracy across systems, as well as standardised data formats that allow integration without manual intervention. On top of this, it means governance frameworks that can underpin both regulatory compliance and AI deployment.
None of this is glamorous. But it is what separates firms that can sustain innovation from those that announce it.
The industry knows where it wants to go. The question is whether the infrastructure underneath can support the journey, and for most firms, the honest answer is not yet.