enGen’s AI Cognitive Engine resolved 2.19 million claim suspensions last year at 99.64 percent accuracy. The architect behind the system says the math was the easy part.
In May 2025, an awards program most people outside the healthcare technology industry have never heard of named enGen’s AI Cognitive Engine the best core administrative processing system of the year. The MedTech Breakthrough panel drew nominations this year from companies in more than 18 countries. The announcement carried on GlobeNewswire on May 8, 2025, ran short. The numbers behind it did not.
ACE, as the engine is branded, addresses more than 3 million claim suspensions a year. Cumulative savings across use cases have exceeded $10 million. The Claims Examiner platform, the system’s adjudication component, resolves 2.19 million suspensions annually at 99.64 percent accuracy, with projected cost savings of roughly $1.07 million per client per year and resolution about 50 percent faster than manual processing. When the AI and a human examiner disagree on a claim, enGen’s published figures show the AI to be the more accurate party 90 percent of the time.
ACE was built inside enGen, the technology operating company of Highmark Health, the Pittsburgh-based integrated health and finance organization that insures roughly 6.9 million members and employs about 44,000 people. enGen itself serves health plans and providers covering more than 20 million members nationwide and employs more than 11,000 people. The named technical figure in the company’s award materials is Rahul Awasthy, the company’s Lead (Principal) Solution Architect. Awasthy holds a Ph.D. in data science, an MBA, and an AWS Solution Architect certification. At enGen, he runs a 25-plus member global team responsible for delivery of the company’s digital health solutions, operating inside a 40-plus member Blueprint team that sets architectural standards across the product ecosystem. He worked alongside teams from Deloitte, Accenture, and the company’s internal engineering organization on ACE’s core architecture, and represented enGen as a technical subject-matter expert at Google Cloud Next 2025.
What he wants people to understand about the award, in interviews, is the part of it that does not appear on the press release. “AI only provides ROI when it is treated as a core business fabric rather than a series of isolated experiments,” Awasthy said.
The framing is a recurring theme in his public talks. It is the explanation he offers for how a system like ACE reached production in an industry that has spent years failing to deploy comparable systems.
In a typical large U.S. payer, claim suspensions are an operational chokepoint. A suspension occurs when an automated claims pipeline cannot adjudicate a claim without human review, usually because of missing data, ambiguous coding, eligibility questions, or coordination-of-benefits issues. Most payers process tens of millions of claims a year and suspend a non-trivial fraction of them. Each suspended claim consumes examiner time, delays provider payment, and degrades the member experience. The industry’s historical response has been to staff the problem with more examiners, more shifts, and more outsourcing.
ACE inverts that response. The Claims Examiner platform applies machine-learning models trained on historical claim resolutions to the suspended-claim queue, classifies each claim against a learned policy framework, and resolves it, in 99.64 percent of cases at the cost target enGen reports, without escalating to a human examiner. The remaining 0.36 percent of cases, plus a structured sample of high-confidence cases routed for audit, escalate to human review. The examiner’s decision is fed back into the system. Model accuracy and audit drift are observable in real time.
The 90 percent figure draws the most reaction in industry forums, and the framing inside the award materials is precise about what it means. The figure describes the population of cases on which the AI and the examiner disagree under the institution’s current operating conditions, and reports that the AI’s resolution holds against the audit ground truth more often than the examiner’s. The distinction is what the governance work depends on.
The MedTech Breakthrough panel evaluates entrants against a fixed rubric across innovation, performance, ease of use, and overall market impact. The combination of accuracy, scale, savings, and operational integration is what the award is designed to recognize, and the panel treated the combination as the qualifying signal.
ACE does not operate in isolation. enGen’s care management platform, branded Predictal, has been reported to reduce prior authorization costs by 75 percent, automate 40 percent of prior authorizations, and produce sub-24-hour turnaround on care access decisions. In Everest Group’s 2025 PEAK Matrix for Care Management Platforms for Payers, enGen was named a Major Contender in its first year of evaluation, among 19 platforms reviewed. First-year-of-evaluation Major Contender placements are unusual. Placements that hold against incumbents on a multi-axis scoring rubric are rarer.
The institutional footprint inside which these systems operate is large. enGen serves health plans and providers covering more than 20 million members. Its parent, Highmark Health, insures 6.9 million. The platforms recognized by MedTech Breakthrough and Everest are running inside that footprint, under audit, with the human-override pathway operating as designed.
Awasthy’s recognition trail at the institutional level reflects the same pattern. He received Highmark Health’s Gold, Silver, and Diamond awards between 2018 and 2022, and an Accenture ACE Award earlier in his career. Those credentials sit alongside, rather than at the center of, the relevant set. What defines the credential set is the recurring presence in the architectural seat at institutions that have to deliver under regulatory load.
The question the industry will be asking for the next 18 months is whether ACE’s architectural pattern is portable. There is no obvious reason it should not be. The pattern, anchored in an explicit governance hierarchy that subordinates sprint cadence to architectural review, is institutionally agnostic in principle. The institutions best positioned to absorb it are the ones with the architectural maturity to run it. That maturity, in U.S. healthcare, is not evenly distributed.
For the moment, the system is processing 3 million suspended claims a year, resolving 2.19 million of them at production grade, and saving an average client roughly $1.07 million annually. The award is the industry’s marker for what is already happening in the footprint. What ACE does next is the more interesting question.
