# Ohio State–Microsoft partnership, reported at $250M, targets 70–80% shift to Azure By MarketScale Newsroom · Published 2026-11-14 · Healthcare on MarketScale Canonical: https://www.marketscale.com/industries/healthcare/ohio-states-reported-250m-microsoft-deal-aims-to-move-70-80-of-computing-to-azure > Ohio State and Microsoft's five-year deal starts with cloud and data work. AI note drafting already runs, and smart hospital rooms come after that. ## Key points - The only part of the Ohio State and Microsoft deal with published results is AI note drafting, which saved up to four minutes per visit in Wexner's 2024 pilot. - Smart-room features such as automatic lighting and infection-risk alerts depend on a unified database that has finished only its high-level design. The first returns are more likely to show up in operating room and staffing forecasts. - Health systems weighing a similar platform deal should ask how the headline value breaks down, which workloads move first and whether the cloud copy of the EHR is a backup or the primary record. Some nurses at Ohio State's Wexner Medical Center spend as much as half their day at a computer on administrative work. CEO Dr. John Warner said so to a room of clinicians and Microsoft staff on Nov. 10, the day the two organizations announced a five-year partnership meant to give some of that time back, according to Samantha Hendrickson's reporting in The Columbus Dispatch. According to Warner, a great deal of new technology has been added, and it has made the work better, with improved outcomes and stronger communication tools. The tradeoff, Warner said, is that part of that gain has come at the cost of time spent with patients and families. The new deal is pitched as a way to win back that time. The Dispatch and Columbus Business First both put the deal at $250 million. As Ohio Tech News pointed out, the joint announcement itself did not disclose financial terms. The two organizations have worked together for years, and the Dispatch describes this agreement as a new endeavor on top of that relationship. It covers clinical AI, a large move of computing into Microsoft's cloud and wider access to Microsoft training and certifications. Coverage of the announcement focused on the smart hospital: rooms that adjust to a patient's condition and alerts that flag infection risk. The announcement describes something else first. It describes several projects moving at different speeds, and most of the money and effort sits in the infrastructure underneath those features. A health system CIO looking at a similar platform deal can learn more from the order of the work than from the headline figure. ## Running, designed and still exploratory Ohio Tech News set out the components and their stages. Lined up together, they show how much of the deal is already in production and how much is still a plan. - Running now: clinicians use Dragon Copilot to draft notes from patient visits, then review and finalize them. A smart hospital platform built on Azure and Internet of Things technology operates at the Richard M. Ross Heart Hospital and the James Outpatient Care facility. - Under way: the medical center is moving its electronic health record to Azure as a backup for business continuity. The partnership targets moving roughly 70% to 80% of Ohio State's storage and computing to Azure, with digital pathology and imaging named specifically. - Designed: a shared data foundation on Microsoft Fabric and OneLake to connect clinical, operational and research data across the medical center, colleges and regional campuses. Its high-level design is complete. - Exploratory: AI-supported tumor-board workflows that would bring pathology, imaging, genomic data and clinical histories into one view for teams reviewing complex cancer cases across locations. The EHR detail is easy to skim past. Azure is getting a backup copy of the health record for continuity, and the announcement does not say the primary system is moving. That makes it a resilience decision. For an operations team, it means the cloud environment has to be ready to carry clinical work during an outage, which is a higher bar than serving as an analytics store. The smart hospital platform at the two facilities connects facility, operational and clinical-support systems. The American Hospital Association, which the Dispatch cited, defines a smart hospital as one that integrates digital tools such as AI to "strengthen clinical efficiency and enhance the patient experience," and describes the cloud-based backbone as an "internet of medical things." Measured against that definition, the Ross and James deployments appear to sit mostly on the efficiency side for now. The announcement says the early work is improving energy efficiency and resource use but gives no figures. That limits how much a buyer can take from the example. Without numbers, the two sites can show a health system which building and clinical-support systems are worth connecting, but not what connecting them saves, so they help with scoping more than with a business case. Ohio State plans to extend the platform to more buildings and to use what it learns in operations at the medical center's new University Hospital. Facilities directors planning new construction now have a real reference site to watch, even though the results are not yet public. ## The one piece with a stopwatch on it Of everything in the Ohio State and Microsoft deal, the part with published numbers is the one that drafts clinical notes, and Wexner's own 2024 pilot put that at up to four minutes saved per visit. Healthcare Finance News editor Jeff Lagasse reported the details at the time. Microsoft's Dragon Ambient eXperience (DAX) Copilot uses conversational, ambient and generative AI to listen to a provider-patient visit and draft notes in the patient's electronic medical record, which the clinician reviews and edits afterward. From mid-January to mid-March 2024, 24 physicians and advanced practice providers in primary care, cardiology and obstetrics and gynecology tested it in outpatient visits. Patients gave permission before each recording. Tripathi said a few clinicians preferred their old workflow, and the pilot group was allowed to keep the tool after the eight weeks of testing. Clinicians caught errors, such as wrong pronouns or one word heard as another, during chart review. On July 1, 2024, Ohio State opened ambient documentation to all outpatient providers, and patient satisfaction scores improved, with patients saying their conversations with physicians were more valuable. That error pattern explains why the review step stays. The partners said their approach is meant to reduce administrative burden while keeping clinical judgment and human oversight central, and the pilot shows what that oversight actually catches: small transcription slips, fixed before a note is signed. Block Field Some simple arithmetic on those figures: 64 hours spread across 100 clinicians comes to roughly 38 minutes each over two weeks. That is modest next to the seven-plus weekly hours of administrative work the American Medical Association has documented, as cited by the Dispatch. It is still a usable baseline. A health system piloting ambient documentation can set its per-visit and per-clinician measures now and compare them with Wexner's. There is a limit to how far these numbers go. They come from outpatient physician visits, while Warner's half-day figure describes nurses, whose screen time comes from different tasks. Note drafting for doctors does not touch most of that. If nurses get relief, it is more likely to come from the connected-data and smart hospital work, which is why the order of the projects matters. ## Why the database comes before the smart room Hether Danforth, Microsoft's vice president for U.S. state and local, told the Dispatch that the first step is a secure, unified database. Departments are working through likely problems before anything goes live, such as matching staff numbers to patient capacity to ease shortages. She said work that patients will notice is coming within the five-year term, and called early governance and privacy work the unglamorous but foundational part. > Often times the beginnings of these projects are a little bit boring because that's what we're working on, the governance to make sure that we can keep out any potential cyber issues., Hether Danforth, Microsoft, to The Columbus Dispatch That emphasis lines up with what the public told Wexner. In a national poll the medical center released alongside its 2024 documentation rollout, 70% of respondents said they were comfortable with AI taking notes during an appointment, yet 70% also had concerns about data privacy and 56% found AI in health care a little scary. For a system joining clinical and research records in one layer, that suggests privacy work is part of earning patient acceptance rather than overhead to rush through. Dr. Ravi Tripathi, an anesthesiologist and the system's chief health informatics officer, gave the Dispatch an example of what that database makes possible. A patient prone to migraines could find the lights already dimmed, without having to ask anyone. That only works if the room's building systems can read something from the patient's record. Without the shared data layer, the smart room is just a room with sensors. The same data could run quieter jobs in the background, the Dispatch reported, such as alerting providers to a patient with a possible safety risk or noting whether a patient prefers a phone call to a message. Tripathi framed the payoff in minutes. He said a medical decision usually has to be made after only a matter of minutes with a patient, and that having the needed information up front would help him earn trust and get patients to share more. Dianne Kennedy, general manager of customer success for Microsoft's health and life sciences business, set an even lower-profile goal: an approach so smooth that patients don't notice it and doctors move easily from one patient interaction to the next. The first uses Ohio State has named for the Fabric and OneLake foundation are operational ones: forecasting surgical demand, staffing needs and operating room use. A secure research environment on Azure could also give investigators controlled access to appropriate, de-identified data. Together this suggests the first measurable returns will show up in OR schedules and staffing plans, with bedside features following once the plumbing is proven. For perioperative and workforce leaders, that means the early work could reach their dashboards before it reaches patients. There is a reason the forecasting work trails the note drafting. Moody's Investors Service said in 2023, as Healthcare Finance News reported, that unlike generative AI, most AI models are not plug-and-play and their integration into business models is expected to take several years. Dragon Copilot, a generative tool, went from pilot to every outpatient provider in about six months. Surgical-demand and staffing forecasts depend on a data foundation that has only reached completed high-level design, and a five-year term gives that slower track room to run. ## Questions to ask before a similar deal For health systems weighing their own multi-year platform agreements, the Ohio State structure points to a few questions worth settling before signing. First, how is the headline value made up: cash, cloud credits, licensing or services, and who pays whom? Ohio State's reported figure and its undisclosed terms show how easily those can get separated in public. Second, which workloads move first, on what schedule and with what size of data estate? A target range like 70% to 80% is more useful once it is tied to specific migration phases. Third, what will count as proof of value? The Congressional Budget Office concluded in March 2024 that evidence on AI's usefulness in health care is mixed, particularly on costs, with models showing usefulness in predicting cancer mortality but falling short on heart failure outcomes. Wexner's per-visit and per-clinician documentation figures are the kind of measure worth writing into a deal before the work starts, rather than reconstructing afterward. Block Field The deal also reaches past the hospital. Ohio State is becoming a Microsoft Training Services Provider and will expand Microsoft Learn resources and certifications for caregivers, researchers, faculty, staff and students, with plans to open training to community members. For central Ohio employers hiring for cloud and data roles, that could matter as much as the hospital work, depending on how community access is set up. Three signals will show whether the infrastructure-first reading holds: the smart hospital platform reaching more Wexner facilities and the new University Hospital, published progress toward the migration target, and the first measured results from the data foundation. If those results are OR-utilization or staffing-forecast numbers rather than bedside features, the deal is running in the order the announcement implies. ## Sources - [Ohio State, Microsoft launch five-year AI and cloud partnership reportedly valued at $250 million](https://www.ohiotechnews.com/ohio-state-microsoft-launch-five-year-ai-cloud-partnership/) - [Microsoft, OSU Medical Center team up to create 'smart hospital' data system](https://www.dispatch.com/story/news/healthcare/2026/11/10/ohio-state-microsoft-ai-smart-hospital-data-system/92706471007/?gnt-cfr=1&gca-cat=p&gca-uir=true&gca-epti=z113501p003750l002050c003750e1159xxv113501d--59--b--59--&gca-ft=116&gca-ds=sophi&ref=ohiotechnews.com) - [Ohio State launching AI app to record patient and physician ...](https://www.healthcarefinancenews.com/news/ohio-state-launching-ai-app-record-patient-and-physician-conversations) (Healthcare Finance News) Tags: Ohio State University, Wexner Medical Center, Microsoft, Microsoft Azure, Dragon Copilot, Microsoft Fabric, smart hospital, healthcare cloud migration, ambient clinical documentation, healthcare IT, hospital operations, IT procurement, clinical AI --- Source: MarketScale, https://www.marketscale.com/industries/healthcare/ohio-states-reported-250m-microsoft-deal-aims-to-move-70-80-of-computing-to-azure. Published for AI indexing and citation; cite the canonical URL. Site guide for agents: https://www.marketscale.com/llms.txt