Why AI matters at this scale

Carelon Global Solutions, the technology-enabled services and business process outsourcing (BPO) arm of Elevance Health, operates at a massive scale, supporting health plans with claims processing, member services, provider data management, and clinical care coordination. With over 10,000 employees, the company manages billions of dollars in administrative healthcare spend annually. At this magnitude, even marginal efficiency gains translate into tens of millions in savings. The healthcare payer sector is burdened by legacy manual processes, complex regulations, and exploding data volumes. AI presents the only viable path to sustainably reduce administrative costs—which can consume 15-25% of premium dollars—while simultaneously improving accuracy, speed, and member/provider satisfaction. For a firm like Carelon, AI adoption is not a speculative innovation but a core operational imperative to maintain competitive advantage and deliver on its promise of simplifying healthcare.

Concrete AI Opportunities with ROI

1. Intelligent Claims Automation: Manual claims review is a primary cost center. Machine learning models can be trained on historical claims data to predict adjudication outcomes, automatically flagging only the exceptions (e.g., potential coding errors, policy violations) for human review. This can increase auto-adjudication rates from ~85% to over 95%, directly reducing labor costs per claim by 30-40% and speeding up payment cycles. The ROI is direct and measurable in full-time-equivalent (FTE) reduction and improved client satisfaction metrics.

2. NLP for Utilization Management: Prior authorization is a notorious bottleneck. Natural Language Processing (NLP) can read and interpret clinical documents (progress notes, lab reports) to extract relevant criteria, automatically comparing them against payer medical policies. This can reduce the manual work for nurses and analysts by up to 70%, cutting authorization decision times from days to minutes. The ROI manifests as reduced labor costs, faster care delivery for members, and improved provider relations.

3. Predictive Provider Network Management: Maintaining accurate provider directories is costly and compliance-critical. AI can continuously scrape, match, and validate provider data from hundreds of sources (claims, credentialing systems, public records) to identify discrepancies, outdated information, and potential network gaps. This proactive approach can reduce member complaints and regulatory fines related to directory inaccuracy by an estimated 50%, protecting revenue and brand reputation.

Deployment Risks for a 10,000+ Enterprise

Deploying AI at this scale introduces unique risks. Data Fragmentation and Quality: Carelon likely operates across multiple client environments, each with its own data schemas and legacy systems. Creating unified, high-quality data pipelines for AI is a massive integration challenge. Regulatory and Compliance Hurdles: Any AI touching Protected Health Information (PHI) must be built and audited within strict HIPAA and potentially state-specific frameworks, slowing development cycles and increasing costs. Change Management: Introducing AI that automates tasks requires significant workforce retraining and role redesign. Managing this transition for thousands of employees without disrupting service levels is a major operational risk. Explainability and Bias: In healthcare, 'black box' AI is untenable. Models denying claims or recommending against authorizations must provide clear, auditable reasons to avoid regulatory backlash and ensure fairness, adding complexity to model development.