Health plans are facing unprecedented challenges driven by rising inflation, high medical loss ratio (MLR), digital disruptions, tough provider negotiations, Medicare Advantage (MA) rate cuts and increased margin pressures, all of which require strategic finance and operations management to respond effectively. In this tumultuous time, traditional business levers are no longer sufficient, and health plans are turning to AI technologies to untap opportunities that can mitigate their challenges by sourcing innovation, scaling productivity and boosting operating income.
According to research conducted by HFS in March 2024, over 64% of health plans believe AI to be a game changer in healthcare. Another study by McKinsey has revealed that health plans could achieve net savings in key areas, like administrative costs by 13% to 25%, medical costs by 5% to 11%, and revenue boost by 3% to 12%, by utilizing AI technology to its best.[3] Health plans have already started experimenting with AI through small pilot projects and proof of concepts, but that is not enough. To sustain, differentiate and grow in a challenging market situation, they need to mobilize and transform their organizations to harness the power of AI to the maximum. This brings us to ask: “How can health plans do more with AI?”
Promising interventions
The surging innovation around AI (natural language understanding, prescriptive analytics, semantic research, content summarization, sentiment analysis, logical reasoning, speech recognition, digital twins and vision technologies) is bringing sizable opportunities for health plans to transform their value chain functions and become increasingly sophisticated and customer-centric.
Many front office activities, such as marketing and outreach campaigns, underwriting, rating and quoting, sales and member and provider communication, can be supported by AI-driven insights and agentic automation, bringing next-generation efficiencies, scale and personalization. These modifications further contribute to improving customer satisfaction, engagement and retention. Similarly, in middle and back-office functions, AI promises to bring operational excellence, efficiency and productivity improvement by automating complex, cost-driven and commoditized administrative processes like enrollment, claims adjudication, prior authorization, appeals and grievance management, network management and provider services, policy and benefit configuration, medical management, compliance tracking and payment integrity. AI further enables health plans to optimize their cost to manage infrastructure, data, software engineering and IT operations lifecycle.
Challenges and cautions
The enthusiasm for any new technology adoption often intersects with many critical concerns, which sometimes hinder the rate and scale of its adoption. Various market studies have revealed that most healthcare organizations, including health plans, are still struggling with their AI adoption across all key dimensions that are considered critical for successfully executing AI-enabled transformation, including business-led AI strategy and roadmap, talent orientation, technology choice, AI deployment and risk management, data ecosystem readiness and AI scaling.
AI models allow health plans to process and analyze vast volumes of data quickly and make more informed and personalized decisions. However, the accuracy of these models depends on the quality and representativeness of the dataset they have been trained on. Historical biases embedded in datasets can cause AI models to hallucinate and reinforce disparities and inequity in care access or incorrectly identify and tag certain high-risk demographic groups. Suppose an AI-led decision engine continues flagging certain members disproportionately for interventions or denies coverage based on incorrect data. In that case, it leads to unintended consequences, including regulatory scrutiny and health plan reputational damage. Investigations and reports from lawmakers suggest that AI algorithmic decision tools can lead to more claims or prior authorization being denied. Regulators are increasingly concerned about AI systems' potential to perpetuate or aggravate biases in healthcare. Ensuring that AI systems are fair, transparent and trustworthy is becoming a key regulatory focus. For example, the Physicians Take Decisions Act, which took effect in California on January 1, 2025, prohibits health insurance providers from using AI as the sole basis for denying claims (AI algorithms were responsible for denying around 25% of all denied claims in 2024).
Another major concern is data privacy and security, given the extensive use of sensitive health and personal information in AI-powered systems. Regulations around healthcare data, such as HIPAA compliance, GDPR and evolving state-level privacy laws, necessitate stringent data management and governance. Any breach or misuse of AI-driven decisions could damage member trust and expose payers to legal and financial risks.
AI models relying on historical data may struggle to keep up with changing regulations, payer policies and unexpected market conditions. For instance, a utilization management system focusing on cost savings might not adapt to new CMS rules or updated clinical best practices, resulting in inappropriate care denials. Therefore, AI models need continuous monitoring, updates and built-in mechanisms to evolve with regulatory and business changes.
The goal of AI adoption in healthcare payers is to drive better outcomes that are both financial and member-centered. However, without a clear focus on measurable improvements, AI can sometimes prioritize efficiency at the expense of member well-being. If an AI-driven utilization management system aggressively reduces costs without considering clinical nuances, it could result in denied claims that negatively impact health outcomes. Similarly, AI-driven member engagement tools must go beyond automation and focus on translating efficiency into improvement in member experience.
Another critical factor here is workforce integration. While AI can reduce administrative tasks, freeing employees to focus on strategic activities, there is often resistance due to fears of job loss and distrust in AI's decision-making processes. Adoption will be challenging if employees don't understand or trust AI's outputs and continue treating it skeptically.
What should be the way forward?
For health plans, embracing AI and ML advancements is crucial. Unlocking the full potential of AI requires more than just adopting another new technology; it demands strategic alignment, strong governance and operational readiness. As health plans evaluate AI investments, balancing innovation with responsibility is key. Successful adoption will integrate AI to enhance transparency in decision-making, improve efficiency and deliver measurable outcomes while mitigating the risks of ethical AI issues.
Overcoming systemic challenges
Overcoming systemic challenges in health plan ecosystems—such as legacy applications, data silos and noisy data—is critical for unlocking AI-driven transformation. Legacy systems often lack interoperability, making unifying data from disparate sources difficult and hindering AI adoption. Inefficient data management, including poor data quality and metadata, further complicates AI model training and performance. To mitigate these challenges, organizations must establish a strong data foundation that integrates disparate sources into a single AI management framework, along with data lineage and observability.
Robust governance and Responsible AI
Widespread AI adoption will also require robust governance controls and Responsible AI guardrails driven by trust, risk and security management (TRiSM) frameworks to ensure data integrity, compliance and transparency. Implementing AI within a controlled environment will prevent biases, enhance explainability and ensure ethical deployment. To create the ultimate ‘Responsible AI Model,’ continuous validation, bias audits and human oversight are crucial for minimizing bias and driving fair and accurate decisions.
Structured use case distillation
Adoption success depends on a structured use case distillation strategy based on business impact, technical feasibility, data availability, user acceptance and cost-benefit analysis. For instance, AI-driven claims automation can significantly reduce processing time and administrative costs (business impact), is technically feasible with large language models (LLMs), relies on structured claims data (data availability) and is highly accepted by payers and providers. Another example is predictive risk scoring for chronic disease management, which integrates clinical and claims data to enable proactive interventions, driving better outcomes for members while being cost-effective. Monitoring AI's impact on operational efficiency and consumer satisfaction through robust performance metrics ensures it delivers real value.
Incremental deployment and scaling
Health plans can prove value by deploying AI in high-impact, low-risk areas with measurable ROI before scaling towards enterprise-wide transformation. As AI-driven insights continuously improve through real-world data, adoption can expand from basic functions to high-value areas, like disease management, precision and personalized care planning. Leadership at health plan organizations must be mindful of the implications of AI adoption on their employees and incorporate effective change management strategies, comprehensive employee training and explainable AI models that offer transparent reasoning behind decisions to increase their trust in AI.
Beyond proof points
Though optimism about adopting AI remains high within health plans, many aren’t taking full advantage of it. They must go beyond the proof points to deploy the technology at scale, mitigating the ethical and legal risks and establishing best practices in data and systems governance. This will help their value chain develop new business dynamics and value propositions that sustain and grow in a competitive market.
References:
- HFS Research -70%+ of Payers and Providers Anticipate Greatest Impact of GenAI on Health Outcomes and Member Experience, According to HFS Research and Cognizant Report. Retrieved March 2024. https://www.hfsresearch.com/press-release/70-of-payers-and-providers-anticipate-greatest-impact-of-gen-ai-on-health-outcomes-and-member-experience-according-to-hfs-research-and-cognizant-report/
- Define Ventures. Inside the C-Suite - Payer & Provider Leaders Share Their Vision for AI. Retrieved March 2024. https://www.definevc.com/insights/inside-the-c-suite-payer-provider-leaders-unveil-their-vision-for-ai
- McKinsey - Taking advantage of the opportunity of AI for payers. Retrieved March 2024. https://www.mckinsey.com/industries/healthcare/our-insights/the-ai-opportunity-how-payers-can-capture-it-now
