AI in the Operating Room: What It Actually Changes
Most perioperative problems are not one person's mistake. They are gaps in what each person can see.
What AI in the operating room changes for surgery centers: scheduling, supplies, first cases and turnover, the numbers to measure, and why clean data comes first.

The perioperative continuum involves more people, more handoffs and more coordination than almost any other part of healthcare operations, which is exactly why AI in the operating room is less about robots and more about information.
Surgical schedulers. Pre-op nurses. Anesthesiologists. Circulating staff. Supply chain coordinators. Recovery teams. Each works with partial information, and each makes decisions that affect the next person in line. When something goes wrong, like a cancellation, a missing supply or an unexpected complication, the cause is rarely one person. It is almost always a gap in visibility. Someone downstream did not get what they needed from upstream, and by the time anyone noticed, the chance to act had passed.
AI-powered orchestration is starting to address this directly. This piece covers what it changes, how to measure it in a surgery center, and what it does not change.
The coordination problem
Most perioperative teams run on experience, institutional memory and manual tracking. Scheduling decisions rest on past patterns and gut feel. Supply use is reviewed after the fact. Staff assignments are handled by people who have done it long enough to know what the system does not tell them.
That works until it does not. Volume rises. Staff turn over. Payer mix shifts. Suddenly the informal knowledge that held everything together is not enough. Late first cases are a good example: the median center starts only about 64 percent of first cases on time (ORBC, via OR Manager), and each late start ripples through the day. See First Case On-Time Starts.
What AI-powered orchestration does is make the implicit explicit. It connects scheduling, supply chain, financial and clinical data into one clear picture. It does not replace the judgment of the people running the operation. It gives them better information to use that judgment on.
How AI in the operating room works in practice
Strip away the marketing and most operating room AI does three things with data the center already produces.
- It predicts. Using each surgeon's actual case history rather than the booked duration, it estimates how long a case will really take, which rooms will run over and which blocks will sit partly empty.
- It watches. It tracks room status, turnover, supply use and staffing as the day happens, instead of reconstructing them from documentation afterward.
- It flags. When something drifts, a first patient not ready, a case cart missing an item, a block likely to go unused, it tells the person who can fix it while there is still time.
None of these is a clinical decision. They are the coordination tasks that currently live in a charge nurse's head and a scheduler's spreadsheet.
Who this changes things for
Every role in the perioperative setting is affected, not because AI takes over tasks, but because better information changes how the work gets done:
- Schedulers see utilization patterns and plan block time more accurately.
- Supply coordinators anticipate demand before cases are pulled.
- Clinical staff spend less time on administrative follow-up and more on patient care.
- Leaders see financial and operational performance together, instead of in separate reports reconciled by hand.
The burden does not disappear. It shifts toward the work that actually needs human judgment. For the full playbook, see Perioperative Optimization: The Hospital Playbook, Adapted for the ASC.
What to measure
If AI is changing the operating room, it shows up in a handful of numbers that already have published benchmarks. Record each one before any tool goes live and check it monthly after.
| Metric | Benchmark | What AI should help with |
|---|---|---|
| First case on-time starts | 90% practical floor; median 64.3% | Readiness flags the evening before and the morning of |
| Turnover time | Under 7 minutes short cases, under 10 complex; survey median 28.5 | Live room status; next patient prepped before close |
| Block utilization | 70% or better; 42% of block time unused on average | Early prediction of unused blocks so time can be released |
| Cases per OR per day | 4.4 on average | Better duration estimates, tighter schedules |
| Cancellations inside 24 hours | About 1% in ASCs | Pre-admission gaps caught days ahead |
| Supply cost per case | 21.6% of collections at the SIS gold standard | Consumption tracked to the case, not the preference card |
Raw utilization can flatter a schedule that runs late into overtime, so many teams also track utilization inside staffed hours; see OR Prime Time Utilization.
A worked example: turnover minutes
The example is illustrative; use your own room counts and case volumes. Take a room running five cases a day, which means four turnovers. At the survey median of 28.5 minutes, those four turnovers take 114 minutes. At a 10-minute standard, they take 40. The difference is 74 minutes a room, every day, which is close to a sixth case in many specialties.
AI does not clean a room faster. What it can do is remove the waiting inside that 28.5 minutes: the next patient not prepped, the case cart not built, the transport not called, the surgeon not paged. Those are information delays, and they are where most turnover time goes. The specialty-level detail is in Turnover Time in the ASC.
What AI does not change
Some fixes are policy, and no software makes them for you. A release-or-lose rule for block time, usually 48 to 72 hours before the date, still has to be agreed by the surgeons and enforced the same way for everyone. A preference card conversation still needs a surgeon, a materials manager and the data on the table. A named owner for first-case readiness is still a person, not a dashboard.
What AI changes is how early those people know, and whether the conversation starts with a guess or a number. Scheduling tools in particular fix visibility, not discipline; the distinction is covered in Surgical Scheduling Software for ASCs.
The foundation under it all
None of this works without clean, connected data. That is the part most AI conversations in healthcare skip. The tools are real and the potential is real. But if the data is fragmented, sitting in disconnected systems, retyped between platforms or living in spreadsheets that are already out of date, AI has nothing to work with.
That foundation is where Sigmatic starts. Our platform collects operational and financial data automatically from the systems where it already lives, uses ambient capture to track rooms, equipment and implants without adding data entry for nurses, organizes it, monitors it continuously and surfaces what matters before a problem becomes a loss or an opportunity slips away. Provider-owners should not need a large internal team to see how their center is really performing. That visibility should come standard. For the benchmarks behind every number in this article, get our free guide, The 2027 ASC State of the Union, or read where AI is already producing returns in AI for Surgery Centers: Where It Is Already Paying Off.
FAQ
How is AI used in the operating room?
In surgery centers, most operating room AI is operational rather than clinical: predicting case durations from each surgeon's history, tracking room status and turnover in real time, matching supply use to the case, and flagging problems like an unprepared first patient or a block likely to go unused.
Can AI improve OR efficiency?
It can, when it is aimed at a measured problem. Record first case on-time starts, turnover time, block utilization and cancellations before starting, then compare monthly against the benchmarks in the table above.
Does AI in the OR replace surgical staff?
No. It shifts administrative follow-up and manual tracking away from clinical staff so they can spend more time with patients. Decisions about care, staffing rules and block policy stay with people.
What data does operating room AI need?
Scheduled and actual case times, room and turnover events, supply and implant use by case, staffing, and ideally billing data so cost and revenue can be seen per case. If that data lives in disconnected systems, connecting it is the first step.
Related reading
Make your ASC more profitable
Sigmatic is a team of ASC consultants who use our own technology to manage surgery centers: better margins, better workflows and a live view of cost per case and utilization. Book a free 30-minute consultation on your own numbers, or grab the free PDF, The 2027 ASC State of the Union, from the form on this page.
Free guide
The 2027 ASC State of the Union: Maximizing profit on razor-thin margins
16 pages of charts: the market, the 2026 reimbursement rules, the benchmarks that separate a 13 percent margin from a 37 percent one, a case costing method you can run this month and a 90-day plan.
Before you go
Want a more profitable ASC?
Book a free 30-minute consultation on your own numbers, or leave your email and we will send the 2027 guide first.
Make your ASC more profitable.
A free 30-minute consultation on your own numbers: where the margin is, what to fix first. No obligation.