(Image credit: IIR/Gemini.)
Software quality pioneer Watts Humphrey once observed that “unplanned process improvement is wishful thinking.” It is an insight from the early days of enterprise computing that resonates powerfully as insurers consider the next evolution in the use of artificial intelligence.
AI, in all its forms, is ultimately a technology like any other. Implemented thoughtfully, it can improve productivity, strengthen decision-making and mitigate organizational risk. When implemented with more hope than thought, it can produce the opposite result: complexity layered onto complexity, cost without corresponding value, and operational fragmentation disguised as innovation.
For many IT professionals, this is not a first rodeo but another chapter in a familiar story. That is not to diminish the transformative potential of AI. Quite the opposite. The issue is less whether AI matters than how organizations choose to operationalize it.
Software quality pioneer Watts Humphrey.
Sustainable improvement rarely comes from isolated experimentation. It emerges through disciplined execution against clearly defined business outcomes. That principle applies across industries, whether insurance, banking, pharmaceuticals or manufacturing. Organizations continuously weigh prospective gains against implementation cost, operational disruption and enterprise risk.
Not every initiative succeeds, nor should every initiative be evaluated against identical benchmarks. Well-managed organizations understand this and approach transformation as a portfolio exercise, extending acceptable risk within defined boundaries rather than wagering the enterprise on a single initiative.
Humphrey’s point was that effective processes cannot simply be willed into existence. One of my finance mentors once observed that surprises, whether positive or negative, are often the consequence of managerial failures.
That observation feels especially relevant in the current AI cycle.
Leadership, Not Magic
Artificial intelligence can produce remarkably good outcomes. It can also produce remarkably bad ones. As several business thinkers have argued in recent years, AI transformation is fundamentally a leadership challenge rather than merely a technology deployment exercise. Peter Drucker and Warren Bennis framed the distinction succinctly: management is doing things right; leadership is doing the right things.
This should serve as a warning to organizations that believe sprinkling AI magic pixie dust across existing operations will somehow produce transformative results. In reality, this is not primarily a technology issue at all. It is a dual challenge involving process and culture. Technology may enable change, but it is not the change itself.
The Risk of ‘AI Theater’
A decade ago, the insurance industry experienced a wave of innovation enthusiasm surrounding the rise of InsurTech and FinTech ecosystems. Important advances emerged from that period, but so did significant misfires. In many cases, organizations underestimated the cultural and operational barriers standing in the way of meaningful transformation. “Innovation theater” became a recognizable phenomenon.
Today, parts of the industry risk drifting into a new version of the same problem: “AI theater.” In some cases, carriers appear to be investing heavily in AI solutions before clearly defining the business problems they are attempting to solve.
In highly regulated industries such as insurance, where decisions made today may carry consequences for decades, governance and disciplined planning are not optional. Experimentation has value, but it must occur within controlled frameworks and with a portfolio view of capital allocation and enterprise risk.
Some studies suggest AI can produce measurable productivity improvements. Others indicate that rework, remediation and quality-control efforts can offset meaningful portions of those gains. In some operational settings, layering AI onto legacy environments risks creating what might be described as “RPA on steroids.” Bandages can prevent infections, but they rarely cure underlying conditions.
Imagining 2036
The real promise of AI lies elsewhere.
Its greatest value is not automating outdated processes more quickly, but fundamentally rethinking how organizations operate. It means imagining 2036 rather than rerunning 1996 in a new wrapper. It means reconsidering the intersection of people, process and data in light of rapidly evolving technological capabilities.
The managerial and cultural implications of this shift cannot be overstated.
The soon-to-be-retired CEO of Coca-Cola recently posed an instructive question: what happens when AI-enabled selling capabilities encounter agentic AI systems operating on the buy side? How must manufacturing, logistics and distribution systems evolve to remain competitive in that environment?
Financial services and insurance organizations should be asking similar questions now.
This is an exciting and consequential moment. New opportunities are emerging alongside new operational and geopolitical risks. AI infrastructure increasingly depends upon massive concentrations of computing power, energy consumption and globally distributed supply chains. Resiliency, concentration risk and long-term operational feasibility will become increasingly important considerations.
Change management has always been difficult. But that is the central challenge facing organizations today.
This is not a CIO punch-list item. It is a C-suite priority.
Game on.