(Image credit: IIR/Gemini.)
AI is making software development faster, cheaper and accessible to more people. It is also making disciplined technology management more important.
For insurers, that creates a difficult paradox. New tools can shorten development cycles and help organizations respond more quickly to changing market needs. But when technology changes faster than strategic priorities can stabilize, speed can become its own source of waste. Teams move rapidly, but not necessarily in a consistent direction.
Leadership Before Technology
One insurance executive recently described an organization in which strategic decisions were made quickly but rarely persisted. Constantly changing priorities consumed resources, created paralysis and prevented promising initiatives from producing sustained value. That is not primarily a technology failure. It is a leadership and management problem.
The same is true of AI-enabled development. Generative AI has given “citizen development” a new meaning: employees can now produce substantial amounts of code without following a traditional software-development path. Yet testing, security review, architecture, documentation and support cannot expand automatically at the same rate. When more code enters the organization than its governance processes can absorb, the result is rework, technical debt and operational risk.
From Buying Software to Building It
Changes in technology economics are adding to the pressure. Cloud computing and software-as-a-service solutions continue to offer valuable speed and scale, but consumption-based pricing can also produce sharp and sometimes unexpected cost increases. At the same time, AI-assisted coding is lowering some of the barriers to internal development. Together, those forces may encourage more insurers to reconsider the balance between buying software and building it.
A recent McKinsey interview with Snowflake’s CEO suggested that AI could effectively turn every company into a software company. The proposition is compelling, but it raises a more important question: What does it mean to operate as a software company?
It does not mean allowing every business unit to develop independently with whatever tools, platforms and architectures it prefers. It means accepting the responsibilities that accompany software creation: managing institutional knowledge, establishing development standards, educating employees, maintaining secure environments and controlling technical debt.
Insurers that have spent years functioning primarily as integration shops may underestimate the organizational change involved. Connecting purchased systems is not the same as creating, maintaining and supporting proprietary software. Early productivity gains can resemble a sugar rush if the supporting ecosystem is not developed at the same time.
New Tools, Familiar Risks
AI coding tools also create new forms of dependency. An insurer may gain speed by standardizing on a proprietary development environment, only to discover later that it has ceded important control to a vendor over which it has limited leverage. The familiar questions of platform lock-in, portability and switching costs have not disappeared. They have simply moved into a new layer of the technology stack.
That is why some practices associated with earlier generations of internal insurance technology organizations may deserve renewed attention. Formal architecture review, source-code management, documentation, quality assurance, security testing, release controls and clear ownership can sound slow or bureaucratic. Properly designed, however, they are what allow development to scale without becoming unmanageable. What is old may be new again.
The need for discipline becomes especially important when development moves faster than an organization can understand emerging risks or anticipate changes in the competitive landscape. Military strategists understand the danger of outrunning supply lines. Insurers face a comparable problem when code production outpaces the people, processes and infrastructure needed to support it.
Accounting for the Shift
The movement from purchasing software to developing it also has financial consequences. CEOs and boards must establish direction, define guardrails and determine how progress will be measured. Technology leaders and finance teams will need a shared view of which costs should be capitalized, which should be expensed and how internally developed assets should be managed over time.
Those decisions can materially affect budgets, financial plans, depreciation schedules and assumptions about useful life. As insurers undertake more internal development, software accounting cannot remain an afterthought. It must become part of the operating model from the beginning.
AI may allow insurers to develop software at unprecedented speed. But speed without governance does not create durable advantage. It creates technical debt, unmanaged costs and systems that become harder to support with every new release.
The insurers that benefit most from AI-enabled development will not be those that generate the most code. They will be those that combine new tools with clear strategic priorities and the disciplined technology-management practices the industry once knew well.