Investment in artificial intelligence (AI) within the healthcare sector is experiencing rapid growth. According to a Forrester report released earlier this year, U.S. healthcare providers are projected to boost their technology budgets to $69 billion this year, with 36% dedicated to AI-driven analytics and software aimed at establishing intelligent healthcare organizations. However, the primary challenge facing these AI initiatives is the rate of adoption, which often lacks a comprehensive strategy.
To achieve the goal of creating an integrated intelligent healthcare organization, many healthcare entities are experimenting with AI pilots and specialized solutions. These add-ons to existing systems will not succeed unless fundamental concerns around data ownership, process design, legacy system issues, and change management are addressed. Leaders in healthcare must prioritize these foundational areas to scale their AI investments for sustainable success.
Transitioning from data ownership to comprehensive data lifecycle management is crucial. The conversation has evolved from merely determining who controls patient health records to considering where data is processed for AI algorithms and how API calls handle sensitive information. Compliance with data residency requirements is paramount to avoid violations when dealing with protected health information (PHI). The choice of deployment—be it cloud, on-premises, or a hybrid model—hinges on these data residency considerations. Infosys, a leader in AI-centric technology consulting, offers over 150 pre-trained healthcare AI models and a wealth of healthcare specialists to help organizations align their cloud infrastructure with regulatory needs.
To tackle data residency issues, a hybrid multicloud model can be effective. This approach maintains the most sensitive clinical data within private cloud settings or local regions while leveraging public cloud resources for less critical data. Additionally, a non-ETL architecture, which allows querying of clinical data without replicating it, minimizes the risk associated with PHI exposure inappropriately.
CIOs may also explore Domain-Specific Language Models (DSLMs) tailored to local medical data, providing accurate, real-time AI insights. Streamlining these models to geographical contexts often yields better contextual performance.
Digital health initiatives have often faltered, primarily due to outdated legacy systems and inefficient manual workflows rather than the technology itself. As organizations consider introducing advanced AI systems, it's essential to first improve the labor-intensive administrative processes. By focusing on optimizing critical tasks such as utilization management and prior authorization, organizations can modernize workflows and implement “human-in-the-loop” touchpoints, which are essential for making these processes AI-capable.
An illustrative case comes from Infosys, which collaborated with a healthcare network to upgrade a legacy case management system, improving compliance and consolidating member data into a unified view. This modernization reduced transaction processing time dramatically, from 70 hours to just 90 minutes, enhancing patient satisfaction by 75%.
As healthcare systems strive to integrate predictive models and AI agents, they must confront the longstanding legacy technical debt that diverts significant IT budgets toward maintenance rather than innovation. While healthcare IT budgets are increasing, a considerable portion is absorbed by the need to maintain outdated systems, posing barriers to true transformation.
Despite this, progress does exist—for instance, modernization of electronic health records (EHRs) has seen considerable strides within the U.S. However, challenges persist, particularly concerning incompatible legacy systems and custom databases that hinder full interoperability.
Healthcare executives must also prioritize organizational change management and prepare their workforce appropriately. Implementing AI tools without ensuring workflows are ready can lead to a "trust tax," where professionals question the reliability and implications of AI in clinical settings. Efforts are being made in leading healthcare organizations to foster discussions around AI use, from concerns about professional autonomy to liability issues. Establishing governance frameworks that emphasize transparency and clinician safety can help mitigate these concerns, leading to a stronger foundation for AI applications.
As the healthcare sector moves towards comprehensive enterprise-wide AI integration, it is essential that leaders focus on modernizing their digital infrastructure first. This foundational step is vital to maximize the benefits of AI for healthcare consumers, the workforce, and overall business operations.
