Scaling AI Without Redesigning theValue Chain Is a Losing Game
True AI impact requires rethinking incentives, workflows, and outcomes—not just scaling technology
By Aneesh Kumar, Chief Digital Innovation Officer, Catalyst Solutions
Artificial intelligence has become a strategic priority for most health plans. Every payer can point to pilots in prior authorization, call centers, claims, or fraud detection. And yet, despite this activity, many organizations are struggling to show meaningful returns. The issue is not a lack of technology. It is a lack of strategy.
Over the past few years, health plans have moved quickly into AI, often outpacing their ability to absorb or scale it. Chatbots replaced IVR systems, point-solution vendors multiplied, and teams ran pilots to test emerging capabilities. That early experimentation was valuable. But many initiatives stalled. Not because the models failed, but because they were often technology-led rather than business-led. Teams focused on what a specific tool or model could do, rather than how a member, provider, or employee experience should work from start to finish. When CFOs asked where the ROI was, the answer was unclear or non-existent.
What emerged across the industry is a familiar pattern: AI efforts that remained siloed, disconnected from core workflow, and unsupported by a unifying operating vision. As vendors, models, and data sources multiply, complexity grows—without clear ownership, accountability, or integration. The result is activity without enterprise impact.
Real AI Impact Starts with Rethinking How the Business Works
The health plans seeing real value from AI are doing something fundamentally different. In-stead of asking, “Where can we deploy AI?” they start with a more basic question: “If technology were no constraint, how should this part of the business actually work?” This shift from tools to outcomes is subtle, but powerful.
When leaders assess processes like member enrollment, provider management, prior authorization, or claims from an end-to-end perspective, it becomes obvious that many pain points are not technical at all. They stem from fragmented workflows, misaligned incentives, and legacy operating models. In that context, AI becomes a catalyst for redesigning how work gets done, not a patch applied to outdated processes.
This broader perspective also forces organizations to examine the full value chain. In areas like prior authorization and fraud detection and prevention, payers and providers often deploy competing AI models, each optimizing locally while driving up system-wide cost and friction. A more effective approach is to rethink how value is shared: how risk is distributed, how outcomes are paid for, how data flows across organizations, and where automation can replace entire steps rather than accelerate inefficient ones. Eliminating unnecessary work creates far more value than automating it.
As AI matures, it is also shifting from task automation to workflow orchestration. Instead of replacing systems, AI increasingly connects them, creating a layer that integrates data, tools, human expertise, and partner inputs into a more coherent ecosystem. The goal isn’t disruption for its own sake. It is a better, more connected operating system that improves accuracy, speed, and experience.
Why Fragmented AI Efforts Become a Drag on the Business
The era of rapid experimentation has run its course. Health plans now have enough evidence, from both successes and failures, to know that isolated pilots create fragmentation rather than compound value. Tools don’t integrate cleanly. Data remains siloed. Compliance risks quietly grow. Vendor sprawl increases complexity and slows operations.
Without a shared strategy, AI becomes a collection of disconnected initiatives that add cost without creating enterprise value.
A coordinated approach does not mean rigid, long-term planning or stifling innovation. It means aligning leadership around a common operating vision, one that clarifies priorities, sets guardrails, and ensures efforts reinforce, rather than undermine, one another. With clarity, teams maintain room to innovate, but within a framework that compounds impact rather than diffusing it.
A Practical AI Roadmap for Health Plans
An effective AI strategy for health plans starts with a few foundational principles:
Business-led ambition
Business leaders, not just IT, should own AI. The goal is to define what the organization wants to become, not chasing the latest model.
Process-first design
AI should reinforce better processes, not preserve old ones. Redesign workflows end to end before introducing AI.
Enterprise alignment
AI impacts talent, operating models, data governance, compliance, and vendor strategy. Treat it as enterprise transformation, not a series of tools.
Domain-aware execution
Healthcare’s regulatory, privacy, and liability landscape is unforgiving. Industry expertise isn’t optional; it is foundational.
Progress over perfection
Waiting for complete certainty guarantees stagnation. Moving forward with a coordinated strategy dramatically increases the likelihood of achieving positive ROI.
Health plans do not need more pilots. They need clarity. Clarity on how the business should operate, where AI genuinely adds value, and how individual efforts connect into a unified operating model.
Without that alignment, AI investments will continue to generate activity without impact—more tools, more vendors, more complexity, but still no enterprise ROI. The plans that act decisively, redesign workflows, and align leadership around a shared vision will be the ones that transform AI from a cost center into a true competitive advantage.
AI is the superpower to connect and orchestrate health plan systems and teams in end-to-end processes. For decades, process experts have been on the sidelines of “real business.” AI brings them to the front, where they start their analysis with industry value chains rather than with internal systems. AI streamlines the integration and orchestration of systems, eliminating the friction and cost traditionally associated with connecting and coordinating complex environments.
The era of ungrounded experimentation is giving way to the era of impact.
