Insurance has never been short of pressure to modernize. What has changed is the pace. AI has moved from a talking point to a tool sitting on almost every desk, and the expectation to do more, faster, is now coming from every direction at once.
That was the backdrop for our recent webinar, where a global panel unpacked what it really takes to transform at scale without losing control of the fundamentals. Here are the themes that stood out.
Why transformation is still difficult in insurance
Insurance is cautious by design, and that caution can slow change. As Melissa put it, one of the biggest hold-ups is often ourselves. The industry leans toward an “if it isn’t broken, don’t fix it” mindset, and a lot of past innovation has been forced on carriers by external events rather than chosen proactively.
There is also the weight of history. Pavan made the point that carriers are not transforming a single process, they are transforming an entire ecosystem built up over decades around legacy platforms, underwriting complexity, claims operations and regulation. Change one area and you feel it everywhere else. Those processes exist for good reasons, which is exactly why unpicking them takes real care.
AI: accelerator, or amplifier of existing problems?
The panel was clear that AI cuts both ways. It genuinely removes friction, letting teams rebuild and scale at a speed that simply was not possible before. But it does not fix a weak foundation. If your data is fragmented or unreliable, AI will not magically clean it up, it will amplify the mess.
Tony summed up a view shared across the group: data is the foundation for any AI ambition. Add a smart layer on top of poor-quality, duplicated or disconnected data and you simply scale the problem. The same applies to process. As Pavan noted, if your workflows are already complex, automating them without addressing the underlying issues just aggravates them.
There was also a shift in where the bottleneck now sits. With engineering sped up dramatically, the constraint has moved to the people directing the work. Speed is only useful when someone is steering it toward the right business outcome.
Data readiness is the real starting line
Ask this panel what “data readiness” means in practice and the answer is trust. It is about establishing a single source of truth that every function across the business can rely on. For insurers running a dozen operating systems, that means standardization, verification and a genuine focus on accuracy.
A few points came through strongly:
- Quality, lineage, ownership and governance are the pillars of readiness. Get these right and models produce better outcomes.
- Timeliness matters as much as accuracy. Many insurers do a strong job preparing data and reporting, but heavy manual preparation and verification slows availability, which erodes the value of the data itself.
- Auditability and control need to run through the whole data process, not just the output.
- Security and appropriate access are non-negotiable. Data can be perfectly clean and still be misused. The right people need the right access at the right time, and no one else.
Legacy systems are not the enemy
One of the more refreshing takes was that legacy does not automatically mean “replace.” If a system is meeting a real business need, and doing it well, there is often little value in ripping it out. The better question is what unmet need you are actually trying to solve.
The panel favored a pragmatic approach. Keep the systems that serve as a reliable source of truth, then build around them where they fall short. If the gap is connectivity, add a layer to bring systems together. If it is data quality, focus on enrichment and accuracy. If it is reporting, close that gap directly. It is usually a more economical route than an expensive rip-and-replace, and it keeps investment focused on the problems that matter.
Rigid roadmaps are giving way to outcome-led planning
The days of neat, multi-year horizon plans are fading. Technology now moves quickly enough that a strategy anchored only to specific tools can be obsolete in weeks. The panel’s answer was to anchor strategy to outcomes rather than technology, then break delivery into smaller increments and revalidate constantly against the business goal.
This looks like continuous horizon and interval planning, with priorities shifting dynamically as capabilities evolve. It means planning holistically rather than betting everything on a single technology, and keeping an organic, human pulse across every line of business even when AI is doing much of the research.
Bringing people with you
Technology change is really people change. The panel described a familiar split in how teams respond: a fixed mindset that sees AI as a threat to hard-won expertise, and a growth mindset that treats it as an opportunity to work smarter and focus on higher-value work.
Setting honest expectations is central to winning trust. Sometimes the most valuable thing a leader can say is that AI is not the answer to this particular problem. There is value in sitting down with users, running demonstrations and proofs of concept, and being upfront that AI features save time but are not perfect, leaving the expert in control to refine the output.
Is this AI moment genuinely different?
The webinar panel thought so. Unlike previous waves such as blockchain, AI has reached across generations and infiltrated everyday life in a way earlier technologies never did. It will reinvent how the industry works, even if the exact shape of that change is still hard to predict. The smart move is to understand clearly what it is good at, what it is not, and how to adopt it deliberately.
Building a culture of resilience
Not every project makes it into production. The consistent message was to celebrate the learning, not just the launch. Discovering early that something is not what customers want, and avoiding a large investment as a result, is itself a win.
Encourage experimentation from the top down, make it safe to fail fast and small, recover, and never miss the lesson. Ideas which do not catch fire the first time often come back around, becoming the foundation for the next build. Progress is rarely a straight line, and that is fine.
The takeaway
Transforming at scale is not about chasing every new tool. It is about getting the foundations right, data you can trust, systems that earn their place, outcome-led plans and teams that are ready to adapt. Do that, and AI becomes a genuine accelerator rather than a shortcut that magnifies existing problems.
At AutoRek, that foundational layer is exactly where we focus. Reliable, controlled and auditable data gives insurers the confidence to innovate on top, without losing control of what matters.