14567
Blog

50 Facilities, 4 Months: Mercy’s Bold AI Deployment Story

aiOS, C-Suite, Foundation Models, Outcomes
Table of Contents

“Every patient that comes to Mercy benefits from Aidoc—big or small. It saves time, money and helps physicians.” — John Mohart, Executive Vice President & COO, Mercy

At Mercy’s leadership summit last fall, Elad Walach, Co-Founder and CEO of Aidoc, joined John Mohart, Executive Vice President and COO of Mercy, for a candid conversation about what enterprise-scale clinical AI deployment actually looks like—and what it takes to get there. What emerged was a story about technology, but more fundamentally, a story about alignment, speed and shared mission.

Platforms as an Accelerator

The conversation opened by framing the central challenge facing health systems today: there are now 1,200 FDA clearances for clinical AI across 700 companies. No organization can realistically manage that many vendor relationships or integrate that many point solutions into coherent clinical workflows. Walach shared that, in Aidoc’s experience working with approximately 150 health system customers, those that pursue a fragmented approach rarely deploy more than three AI tools at scale. Not for lack of intent, but because the integration burden makes it nearly impossible.

Mercy chose a different path. Mohart described their deliberate decision to pursue standardization and a platform approach. They did this by building infrastructure and guardrails that allow AI to scale across every facility, rural and large alike, with the same standard of care. When Mercy and Aidoc first connected in September 2024, the goal was enterprise-wide deployment in months, not years. By February 1, 2025, Aidoc’s aiOS™ was live across all 50 Mercy facilities, running over a dozen use cases simultaneously, scaling to all communities in Mercy in just 30 days.

“It was indeed one of the fastest implementations of this scale we’ve ever seen,” Walach said. “There was not one solution. It was over a dozen use cases all at once, enterprise wide. And this was because of a truly unique meshing of leadership and alignment.”

Early results at Mercy reflect the impact and scale of that commitment:

▸  More than 2.4 million images analyzed by AI in the last year

▸  More than 249,000 studies flagged with critical or actionable suspected findings in the last year

▸  90% reduction in time-to-diagnosis for outpatients with suspected critical findings

▸  Radiologists describe the workflow change as augmenting—not disrupting—their practice

Reimagining clinician and patient experiences

Mohart brought the impact to life with examples from Mercy’s own patient population. For example, when a patient arrives Friday afternoon for a routine outpatient scan, and the study enters a queue that won’t be read until Monday, a pulmonary embolism (PE) discovered incidentally—present in two to four percent of cancer patient scans—can wait days before being read by a radiologist. Aidoc’s aiOS™ flags these patients in real time and notifies the radiologist immediately. Mercy has also used this capability to triage patients in waiting rooms with head bleeds and pulmonary emboli who would otherwise have waited for care.

Walach noted that the impact compounds when data is synthesized across imaging AI and the electronic health record (EHR). As an example, for PE cases, Aidoc can surface data from  CT scans and EHR records – such as troponin levels, right ventricular measurements, oxygen saturation and echo findings – supporting efficient communication across care teams in   coordinating patient care.

“The technology is in the backseat position,” Walach said. “It’s all about making physicians better and helping them deliver the care they believe their patients need as quickly as possible. ”

What’s Next: Foundation Models and a Broader Application

Walach outlined what he called the most significant technological shift in clinical AI to date: foundation models. Where traditional algorithms are trained to identify a single disease state, foundation models are trained on massive, diverse datasets and can identify patterns across hundreds of conditions simultaneously. In partnership with technology innovators NVIDIA and Amazon, Aidoc has invested more than $300 million in training a foundation model,  tackling abdominal and chest imaging as its first solutions. Early results show 99% sensitivity and specificity—compared to roughly 90% for traditionally-built algorithms—with a tenfold reduction in false positives and false negatives across more than 15 disease states. Aidoc has already received two FDA clearances built on the model.

Mohart pointed to calcium scoring as one new near-term application. Calcium scoring is among the strongest predictors of cardiovascular disease risk, but most patients never receive it as a dedicated test. With AI, Mercy can identify patients with elevated calcium scores from chest imaging they are already undergoing—no additional scan required. Based on Mercy’s imaging volumes, Walach estimated 30,000 to 40,000 patients could be surfaced, with EHR data used to prioritize those not already under active cardiovascular management.

At the close, Mohart and Walach aligned on a shared view: AI does not replace the physician. It augments the physician to give every patient the benefit of everything medicine already knows, surfaced at the right moment, for the right care team. The partnership between Aidoc and Mercy demonstrates that powerful AI-Physician collaboration in practice.

Explore the Latest AI Insights, Trends and Research

Follow a manual added link

Webinars

Aidoc experts, customers and industry leaders share the latest in AI benefits and adoption.

Follow a manual added link

eBooks

Explore how clinical AI can transform your health system with insights rooted in real-world experiences.

Follow a manual added link

AI Strategy

Learn how to go beyond the algorithm to develop a scalable AI strategy and implementation plan.

Follow a manual added link

Solutions

Explore how Aidoc can help increase hospital efficiency, improve outcomes and demonstrate ROI.

© Copyright 2026 Aidoc