Highmark Is Hiring for AI Scale Amid Utilization Pressure
Job postings reveal how AI is being embedded across claims, data, and workflows
Highmark Health is approaching 2026 under constraints it could not fully price away. Through the first nine months of 2025, the Pittsburgh-based nonprofit reported $24.6 billion in revenue alongside a $69 million net loss and a $204 million operating loss. Pressure was concentrated in the insurance business, where utilization failed to normalize as expected and remained elevated throughout the year.
“For 2025, we thought the trend would subside and come down just a little bit. They remained elevated throughout the year,” Carl Daley, chief financial officer and treasurer at Highmark Health, told Fierce Healthcare. “Pricing in the first half of the year did not fully contemplate the trends that we’re seeing.”
During the same earnings discussion, executives cited technology and artificial intelligence as factors already influencing claims operations. Kate Musler, chief financial officer of Highmark Health Plan, said the insurer was seeing AI improve provider coding accuracy, easing some cost pressure even as utilization remained high.
Recent AI hiring activity at Highmark aligns with that operating context, with roles focused on enterprise systems.
Turning Internal AI Into Production Systems
Highmark’s use of artificial intelligence extends back more than a decade through machine learning and analytics in administrative functions, with generative AI introduced internally in 2023.
“We’re really trying to make every single one of the Highmark Health employees not only a user of, but a builder of AI,” Richard Clarke, Highmark’s chief analytics officer, said to Pittsburgh Post-Gazette.
In early 2024, Highmark launched Sidekick, an internal generative AI sandbox. By late 2025, the tool had grown to more than 13,000 monthly users and logged roughly two million prompts. Leadership described the system as a controlled environment for testing use cases and building familiarity with generative tools.
Patient-facing AI deployments followed. Highmark rolled out ambient clinical documentation from Abridge across locations operated by Allegheny Health Network, allowing physician-patient conversations to be automatically summarized into clinical notes. Clarke said opt-out rates were in the single digits across tens of thousands of patients offered the technology.
Highmark also outlined plans to expand Care.ai, using computer vision in medical centers to detect fall risk and pressure ulcers, and to integrate AI into call center operations.
Each of these deployments targets documentation, monitoring, or administrative throughput, areas that scale directly with utilization and staffing intensity.
Hiring for Scale Under Margin Pressure
Highmark’s current AI-related job postings focus on platform architecture, data infrastructure, and operational integration. Roles include a principal technical architect for AI platforms and services, a senior AI consultant responsible for proof-of-concept translation and adoption, and a principal data architect overseeing data movement, enrichment, and orchestration across a hybrid enterprise environment.
The descriptions emphasize enterprise integration, governance, and production reliability rather than model research or exploratory development. Responsibilities include formal architecture review, modernization of legacy systems, and deployment of AI services across heterogeneous environments.
That hiring profile aligns with how executives have discussed AI’s role in current operations. Musler’s comments placed AI directly inside the claims and reimbursement process, while Clarke has linked AI deployments to reductions in administrative workload rather than clinical decision-making.
Elevated utilization increases the volume and complexity of claims, documentation, and care coordination, placing additional load on administrative systems.
Highmark’s recent AI hiring show the systems required to operate under those conditions. Utilization remains elevated, margins remain under pressure, and the company is expanding the architecture needed to support AI as a core part of daily operations.