Insurance is unbundling. Work is fragmenting across partners, platforms, and specialized functions. But most insurers are still running it as if it isn’t.
Across the industry, distribution increasingly sits with partners. Product innovation is no longer confined within the enterprise. Operations are spread across internal teams and third parties. At the same time, AI is accelerating decisions closer to the various points of action.
And yet, inside most organizations, decision-making still sits where it always has.
That gap is where transformation begins to fail.
I explored how AI exposes the limits of today’s insurance operating models in a previous piece for Canadian Underwriter. It’s becoming clearer that even when organizations recognize this, progress is uneven and often stalls.
The issue is not capability. It is design.
Insurance is not one business
For decades, insurers operated as vertically integrated organizations: underwriting, sales, claims and risk management were all housed under one roof. That model made sense when coordination was expensive and scale provided advantage.
That environment no longer exists.
Today, what appears to be a single business is, in reality, a combination of fundamentally different economic models operating under one structure. There are customer relationship businesses that win on distribution and access, product innovation engines that compete on speed and specialization, and risk and infrastructure platforms that rely on capital discipline.
Each of these operates differently. Each requires distinct capabilities. Each moves at a different pace. Yet most insurers continue to manage them as one.
What looks like a single company is, in practice, multiple businesses forced into a single operating model.
The value chain is unbundling — but not uniformly
The traditional value chain is no longer tightly integrated. It is becoming more modular.
Customer ownership is shifting. Product development is fragmenting. Service is becoming industrialized. Claims is evolving into a real-time decision system. Capital remains central, but increasingly data-driven.
The industry is moving from integrated firms to a more distributed system. In response, insurers are making sharper choices about where they can truly win. In most cases, that means being distinctive in one or two areas and relying on partners, platforms, or external capabilities for the rest.
What matters, however, is that this shift is not uniform.
Also in the news: Recovery | Injury claims will reveal limitations of Ontario auto reforms
Certain layers, particularly those tied to capital, risk and long-duration liabilities, require control, consistency and centralized governance. Fragmenting these decisions does not create advantage. It introduces risk.
The emerging operating model is not simply distributed. It is selectively designed, with different parts of the value chain governed in fundamentally different ways.
The enterprise has not caught up
What you see in practice is a growing disconnect between how the business operates and how it is designed.
Unbundling has occurred at the level of execution. It has not been matched by a corresponding shift in decision-making. The result is familiar. Decisions take longer. Accountability is blurred. Capabilities are duplicated. Friction emerges across internal teams and external partners.
These are not execution issues. They are design issues.
What is becoming increasingly clear is that this is no longer just an operational problem. It is an economic one.
As AI accelerates quoting, underwriting, servicing and claims in parts of the value chain, organizations with slow approvals and unclear ownership are not simply less efficient. They are less able to capture value.
Parts of the system are already accelerating. Others are creating measurable drag. That gap is now visible at the leadership level, where time and energy are increasingly absorbed by diagnosing misalignment, latency and strain rather than moving the business forward.
The real shift is decisional
This is where most transformation efforts quietly break.
Each part of the value chain operates differently. Yet many insurers continue to apply a single operating model and a single decision framework across all of it.
The real shift is not structural. It is decisional.
Decisions are moving closer to the customer, into workflows, and across organizational boundaries. But decision ownership has not followed.
The value chain is unbundling operationally, while decision ownership remains bundled.
As AI accelerates performance, the cost of that misalignment increases materially. The gap is no longer between organizations experimenting with AI and those that are not. It is between those that can absorb faster, more distributed decision-making, and those that cannot because decision ownership, operating model and execution are no longer aligned.
Unbundling without redesign creates complexity
Unbundling itself is not the problem. In many cases, it is necessary.
The problem is unbundling without redesign.
I have seen this play out repeatedly. External partners execute work that still requires internal validation. The same decisions are made in multiple places, often with different standards. Guardrails are inconsistent. Escalation paths are unclear.
Execution becomes distributed. Decision-making does not.
The result is not agility. It is fragmentation.
Unbundling does not simplify the system. It exposes it.
This is the point many organizations miss. They redesign workflows, introduce partners and invest in technology, but they do not redesign how decisions actually get made across that system.
Start with the map, not the technology
Most transformation efforts begin in the wrong place.
They start with technology, capability builds, or AI use cases. A more effective starting point is to map the value chain end-to-end and make decision ownership explicit.
This is not a conceptual exercise. It is a practical one. Where is value created? Where are decisions made? Where should they be made?
That is where bottlenecks, duplication, and delay become visible.
Without that clarity, organizations do not redesign the enterprise. They digitize its flaws.
AI is already creating uneven economic value
AI is already creating measurable impact across the insurance value chain. It is improving targeting and cross-sell, reducing churn, accelerating underwriting and compressing decision timelines.
But that value is not evenly distributed.
AI is not creating uniform advantage. It is amplifying differences between parts of the system that are able to move quickly and those that are not.
The question is no longer where AI can be applied. It is where it is already creating economic advantage, and whether the enterprise is designed to capture it.
The missing layer: Decision architecture and governance
This is the gap.
Not more technology. Not more partners. Decision architecture.
Designing where decisions sit, how they are made and how they are governed across a distributed value chain has become a core leadership challenge.
Some decisions must remain centralized, particularly those tied to capital, risk and regulatory accountability. Others belong within domains such as underwriting rules, or claims frameworks. A growing share must sit at the edge, embedded directly within workflows, where speed and scale matter most.
The discipline is not about centralization or decentralization. It is about deliberate placement.
Where decisions sit, however, is only part of the equation. How they operate matters just as much.
Most governance models are still built around reviewing individual decisions. That approach does not scale in a distributed system.
What is required instead is system-level governance. Clear rules, defined thresholds, consistent data inputs, and explicit accountability for outcomes.
Governance is no longer about approving decisions. It is about designing and overseeing the system that produces them.
AI accelerates, but does not solve
AI enables faster decisions. It does not define them.
Without a clear decision architecture, organizations risk scaling inconsistency rather than eliminating it.
AI is already creating value. The gap is whether organizations are structured to capture it.
This becomes most visible in where value is actually being created. It will not come from efficiency alone, but from better risk selection, faster product innovation, stronger claims outcomes, and distribution advantage.
The winners will not be the most efficient operators. They will be the ones that align decision ownership to where value is created.
The real work of transformation
The industry is not short on capability.
It is short on coherence.
The question is no longer whether the value chain will continue to unbundle. It already has. The question now is how deliberately organizations choose to operate across it.
This is where the opportunity sits.
Successfully transforming insurers will not simply invest more in technology or expand their partner ecosystems. They will take a more deliberate approach to how decisions are designed, placed and governed across the enterprise.
In doing so, they will unlock something different. Faster execution without losing control. Distributed models that still operate coherently. The ability to absorb AI-driven change without creating additional complexity.
In this environment, delay is no longer neutral. It compounds value leakage across the chain. But the inverse is also true: when decision ownership is clear and aligned to where value is created, speed and performance reinforce each other.
Successful organizations will not move the fastest. They will design their systems to move with clarity, consistency, and intent.