Recent industry reports have highlighted the scale of this challenge, pointing to a growing disconnect between the insurance sector’s investment in AI and the returns being generated from the technology. While AI spending as a proportion of revenue is projected to triple in 2026, just 38% of property and casualty insurers are currently generating value from AI at scale across their core workflows.
As insurers continue to expand their use of the technology, the findings raise an important question: are organisations investing in AI faster than they are preparing the information foundations needed to make it successful?
The problem isn’t the sophistication of the technology itself. Instead, it’s information environment underpinning AI that is disconnected, ungoverned, and missing context.
Insurers, like many other industries, are investing a lot in AI. And big investments require big returns. There’s a lot of focus on what these new technologies will do for their firm but, perhaps, not enough focus on if they’re ready to start using them. That means figuring out if information AI will work with and depend on is properly governed, and has the context needed.
Policy, claims, and compliance information scattered across different systems, poorly classified or disconnected from the processes they relates to means even the most sophisticated AI will be working with an incomplete picture. Rubbish in, rubbish out. It is vital to fill the gaps before embarking on large-scale AI transformation to make the most of this new technology.
For an industry built around assessing risk and making information-intensive decisions, this represents a significant challenge. Policy documents, claims information, customer records, and compliance content can remain fragmented across different systems and departments, limiting the ability of AI to access the complete picture needed to produce reliable outcomes.
Becoming AI-ready, therefore, requires more than selecting the right technology. Insurers need to understand the current state of their information, identify where critical knowledge is fragmented, inaccessible or poorly governed, and address those weaknesses before AI is scaled across business-critical processes.
AI readiness starts with knowing where your business is, before plotting a course on where you want to get to. Is information accurate, accessible, properly governed? Is it connected to the business context that gives it meaning?
If a firm can’t be sure of this, then adding AI on top won’t fix those underlying problems. In many cases, it will simply expose, and even amplify them.
The consequences extend beyond disappointing productivity gains. Underwriting claims and compliance processes all depend on accurate and relevant information. An AI system may be capable of retrieving a document, but that does not necessarily mean it understands whether it is current, how it relates to a particular customer or policy, or whether other information changes its meaning.
This becomes increasingly important as insurers move beyond experimentation and begin embedding AI across wider business processes. Technology that performs effectively in a controlled environment can face very different challenges when deployed across organisations where information is distributed across legacy systems and departmental repositories.
Firms may interpret disappointing AI results as a technology or tool problem when the reality is that the problem might lie in the content the tools are consuming. Insurers need to understand these gaps before they scale, otherwise they risk investing in a technology that isn’t set up for success.
Rather than treating information management as a separate transformation project, insurers should consider it a fundamental part of AI readiness. Understanding where information sits, identifying gaps in governance and accessibility, and connecting knowledge to the customers, policies, claims and processes it relates to, can create a much stronger foundation for AI adoption.
The insurance industry has a massive opportunity to improve decision-making, productivity, and customer outcomes with AI, but this new technology won’t deliver that transformation or value on its own.
Firms that understand the state of their information, address the gaps and build the right foundations before building new tools and technology on top will realise AI’s potential faster than those that don’t.
(Yohan Lobo is Industry Solution Manager for Financial Services at M-Files)