Healthcare AI agent · Perioperative
Surgical-Scheduling & OR-Utilization AI Agent
Post cases against block time, room, equipment and staffing rules, estimate duration from your own history and chase what each case needs — clinical priority and any bump stay with the surgeon and anaesthesia.
- Typical delivery
- 4–6 weeks
- Deployment
- Your stack
- Case priority
- Surgeon only
- After launch
- Agent Care
What this agent does
Fills block time, does not rank cases
In
- 01
Take postings, add-on requests, cancellations and block releases from OR scheduling, the offices and pre-op.
- 02
Read the block grid, room and equipment constraints, preference cards, roster and pre-op requirement lists.
Reason
- 03
Estimate case duration from this surgeon's own history for this procedure, and carry how wide that estimate is.
- 04
Check the room, trays, implants, vendor cover and the anaesthesia and nursing staffing the case needs that day.
- 05
Check clearance, anaesthesia review, authorisation and patient instructions against the requirement list the case carries.
Decide
- 06
Detect a case that will not fit the block, a requirement not met or a constraint it cannot meet — hold for a person.
- 07
Detect anything that would move, bump or reorder a case, and route it to surgeon, anaesthesia and charge nurse.
Out
- 08
Post, move and release inside the rules, then publish the day's list, the open time and what is outstanding.
- 09
Retain the estimate and its basis, the constraint applied, the release decision and every scheduler change.
Product statement
The agent proposes a schedule inside rules a person set. Clinical priority, case order and any bump stay with the surgeon, anaesthesia and the OR charge nurse.
Example workflow
One case posting, end to end
- 1Posting receivedBooking request, add-on, cancellation or a block release from a surgeon's office
- 2Rules readBlock grid, room and equipment constraints, preference card, staffing roster and requirement list
- 3Duration estimatedThis surgeon's own history for this procedure, with the spread the estimate carries
- 4Constraints checkedRoom, trays, implants, vendor cover, staffing, clearance and authorisation — each in place or named as missing
No human action required
Stages 1 to 4 run without a person in the loop — the rules are read, the duration estimated and the constraints checked before anyone is asked to look. Anything that would move a booked case ends that stretch on the spot.
Posts on the approved path.
Goes to the OR scheduler, or to the surgeon and anaesthesia.
Scheduler or charge nurse
Sees the estimate and what it was built from, the constraint that blocked the posting and what is still missing, then posts it, overrides the constraint or takes it to the surgeon.
Post · Override the constraint · Take it to the surgeon
Posted — handed back
- 6Posted to the scheduleWritten only where write access and scheduling policy allow it; no booked case is moved
- 7Day reconciled and evaluatedActual against estimated, on-time starts, turnover, day-of cancellations and released block, by service line
Corrections
Cases a scheduler re-times or moves by hand are counted in the evaluation.
What should not run autonomously
Human approval stays in control
Outside the boundary — human approval required
8 items
- Deciding which case takes priority over another.
- Bumping, moving or reordering a booked case.
- Setting or changing a case's clinical urgency.
- Booking into another surgeon's block without release.
Automation boundary
Agent acts unaided
- ✓Post, move and confirm cases inside the block, room and equipment rules.
- ✓Estimate case duration from your own history and show how wide it is.
- ✓Chase clearance, authorisation, trays, implants, vendor cover and staffing.
- ✓Release unused block on the agreed clock and report what came back.
Write actions run only inside the approval boundaries agreed during implementation. Clinical priority and bumping are not among them.
- Overriding a staffing or anaesthesia coverage rule.
- Posting a case before clearance or authorisation.
- Changing block allocation, release clocks or targets.
- Waiving an instrument, implant or vendor hold.
Example output
One case posting, annotated
Everything the agent proposes is attached to the posting and the rules it was read against.
Scheduling output · single posting
Illustrative example- Posting
- Robotic case, Tuesday block
- Surgeon booked
- 120 min
- Block remaining
- 135 min
- Estimated duration
- 155 min
- Confidence
- 91%
- Case priority
- As the surgeon set it
As received
The posting, the time the surgeon asked for and the block as it stands, kept as sent — nothing on this side is re-decided.
Evidence used
- Prior cases, same surgeon
- Preference-card tray status
- Anaesthesia cover that day
Why it did not post
The estimate runs past what is left in the block by more than the day can absorb, so the posting is held.
What the score decides
Confidence decides whether a scheduler checks the posting, not which case goes first. The surgeon decides that.
Action
PostOverride the constraintTake it to the surgeonValue
Where AI adds value
The same four claims, placed at the point in the workflow where each one applies.
Where the value lands
Value 01 – 04
Every posting and add-on
OR scheduling, surgeons' offices and pre-op
Apply the client's own scheduling rules
Use the block grid, room and equipment constraints, preference cards, staffing roster and requirement lists already on file.
Take the chasing off the OR schedulers
Durations, trays, implants, vendor cover, clearance and authorisation are worked without someone holding a list and a phone.
Put the conflicts in front of a person
A case that will not fit, a requirement not yet met and anything that would move a booked case reach a person before the day.
Build an evidence trail
Retain the estimate and its basis, the constraint applied, the release decision, the chasing done and the scheduler's correction — on both paths.
Integrations
Typical integrations
Five system groups connect to the same agent. Which of them are in scope is decided in discovery.
OR scheduling & block
- Epic OpTime
- Oracle Health
- MEDITECH
- Block grid
- release and add-on rules
Anaesthesia & staffing
- Anaesthesia record
- pre-op clinic
- Rostering
- call and cover rules
Instruments & implants
- Preference cards
- sterile processing
- Implant, tray and vendor booking
Agent
Surgical scheduling & OR utilisation
- Reads the rules
- Estimates and fits
- Holds the conflicts
Pre-op & authorisation
- Clearance status
- anaesthesia review
- Authorisation
- patient instructions
Analytics & evaluation
- Utilisation reporting
- OpenTelemetry
- Langfuse
- Supported monitoring sources
Integration availability depends on the client's existing systems and API access.
Agent controls
Six layers between the model and your schedule
Each control wraps the one inside it. A posting clears every layer before it reaches the schedule, and clinical priority sits outside all six.
L6 · Outermost — last line of defence
Inward → L1 · closest to the model
- L1 – L2
- Decide whether the case may be posted
- L3
- Decides what may reach the day of surgery
- L4 – L5
- Keep priority with the surgeon and the trail intact
- L6
- Pulls automation back when signals degrade
How Nestack evaluates it
Evaluate the whole day — not only how full the rooms were.
Coverage runs the whole depth of the workflow, and every layer is cut by slice.
Surface — the schedule the team is given
- E1Final-output evaluationWas the room, the block placement and the duration right?
- E2Step-level evaluationDid it read the right block rule, card and requirement list?
- E3Tool evaluationDid it read and write the correct case, room and block?
- E4Duration calibrationDo the estimates hold against what the cases actually took?
- E5Slice evaluationHow does scheduling quality change across service lines and cohorts?
- E6Business outcomeHow much was cancelled, delayed, bumped or released too late?
Floor — the case that runs when it was planned to
Failure modes
Where each failure originates in the agent
Seven failure modes plotted against the five stages of the agent lifecycle.
Agent lifecycle
Direction of processing →
01 · Posting / retrieval
2 modes
- SS-01
Preference card out of date
The set pulled is not what the surgeon now uses.
- SS-02
Pre-op requirement missed
Clearance or authorisation is missing on the day.
Stage gathers
The posting, block rules, cards, roster and pre-op
02 · Estimate & fit
1 mode
- SS-03
Duration underestimated
The block overruns and the last case is cancelled.
Stage proposes
Duration, room, block placement and how wide it is
03 · Schedule / write
2 modes
- SS-04
Posted without confirmed cover
Staffing was assumed, not checked against the roster.
- SS-05
Booked case bumped
An add-on is posted over a case already booked.
Stage posts
Only where write access and scheduling policy allow it
04 · Chase & release
1 mode
- SS-06
Block released too late
Time comes back when nobody can still use it.
Stage sends
Requirement chasing, block release and the day's list
05 · Change / Version
1 mode
- SS-07
Utilisation narrows access
The same service lines keep losing the open time.
Stage tracks
Model, prompt, block-rule and release-clock changes
- Sev-1 · acted outside the boundary
- Sev-2 · the case is delayed or cancelled
- Sev-3 · capacity is wasted or access narrows
Affected slices
Overruns and cancellations do not fall evenly
Aggregate utilisation can look healthy while a small number of case cohorts carry most of the overruns, most of the day-of cancellations and most of the bumping. Nestack reports performance by slice, not only in total.
Slice performance — reported separately, not only in aggregate
Illustrative example| Slice | Failure rate | Lift | Lift vs. threshold | Status |
|---|---|---|---|---|
| Long, high-variance procedures | 5.9% | 3.7× | Review | |
| Add-on and emergency insertion | 4.4% | 2.8× | Review | |
| Surgeons without regular block | 3.1% | 1.9× | Watch | |
| Routine short cases | 1.0% | 0.6× | Normal |
Bar: lift vs. routine short-case baseline · scale 0–4.0× · tick marks the 2.0× review threshold
2 of 4 slices over threshold
Evidence-linked improvement
Every estimate you change moves someone else's case
An overrun is not closed by re-timing one case. The fix lands in a duration basis, a constraint rule or a release clock — and each decides which cases fit.
Improvement cycle · five stages
Switchback — the path turns at Improve and returns at Learn
- 01
Detect
Overruns, day-of cancellations or bumping move in a cohort.
- 02
Diagnose
Traced to the estimate, the constraint read, the chase or the release clock.
- 03
Improve
The duration basis, constraint rule or release clock is re-approved and version-linked.
- 04
Verify
Re-run against held-out postings from the affected cohort, including the ones that failed.
- 05
Learn
That case becomes a regression case and the changed rule enters the block policy.
Learn → Detect
The return edge. A release clock or a duration basis is an access decision, so the next cycle is read by service line and by who was bumped, not only by utilisation.
Typical build scope
Twelve workstreams across six weeks
The build scope read against the delivery timeline. Week structure follows the six-week plan — discovery, block rules and durations, constraints and chasing, evaluation, integration, then production validation and handover.
Workstream
Week 1Week 2Week 3Week 4Week 5Week 6
- 01Workflow discovery and automation-boundary definition.Week 1
- 02Scheduling and block APIs, and access scope.Week 1, Week 2
- 03Block grid, release clock and add-on rules.Week 1, Week 2
- 04Case-duration basis from your own history.Week 2, Week 3
- 05Room, equipment and staffing constraints.Week 2, Week 3
- 06Preference-card and implant checks.Week 3
- 07Pre-op clearance and authorisation lists.Week 3, Week 4
- 08Bump, add-on and escalation routing rules.Week 4
- 09Evaluation suite, parity slices and regression cases.Week 4, Week 5
- 10Utilisation, release and day-of reporting.Week 5
- 11Scheduling write-back and reconciliation.Week 5, Week 6
- 12Observability, deployment and Agent Care handover.Week 6
12 workstreams · 6 weeks · bar shows the weeks a workstream is active — several run in parallel
Final scope and sequence confirmed in discovery
Engagement tiers
What each tier includes
Rows are the capabilities named in each tier's scope. Higher tiers include everything below them.
| Capability✓ in scope · — not at this tier | PilotOne service line | ProductionProduction OR integration | AdvancedMulti-site / multi-service |
|---|---|---|---|
| Introduced at Pilot | |||
| Rule-checked case posting | ✓ | ✓ | ✓ |
| Block, room and equipment rules | ✓ | ✓ | ✓ |
| Surgeon-owned priority and case order | ✓ | ✓ | ✓ |
| Baseline evaluation | ✓ | ✓ | ✓ |
| Introduced at Production | |||
| Duration estimates from your history | — | ✓ | ✓ |
| Pre-op, clearance and authorisation chasing | — | ✓ | ✓ |
| Preference-card and implant checks | — | ✓ | ✓ |
| Block release on the agreed clock | — | ✓ | ✓ |
| Observability and evaluation | — | ✓ | ✓ |
| Introduced at Advanced | |||
| Add-on and emergency routing controls | — | — | ✓ |
| Utilisation reporting with access parity | — | — | ✓ |
| Build price | From $5,000 | From $8,000 | Custom quote |
- Final build price
- Confirmed after discovery based on scheduling and block-management integrations, service-line complexity, case volume, staffing and materials systems, minimum-necessary access scope, approval controls and deployment requirements.
- Separate from build
- Build pricing is separate from recurring Agent Care, which covers managed monitoring, evaluations, incidents and verified improvements after launch.
What we need from you
What you bring, and what we build with it
Each input maps to a piece of build scope and a week in the delivery timeline.
You bring
We build with it
01Your block grid, release clock and add-on rules
→Block, release and add-on rule configuration
Week 102Case history for the procedures in scope
→Case-duration basis built from your own history
Week 203Access to OR scheduling, block and staffing APIs
→API assessment, access scope and integration setup
Week 204Room, equipment, tray and staffing constraints
→Room, equipment and staffing constraint model
Week 305Preference cards and what each case needs beforehand
→Preference-card, implant and pre-op requirement checks
Week 306Cases that overran, were bumped or cancelled on the day
→Evaluation suite, parity slices and regression cases
Week 407Named schedulers, and who owns priority and bumping
→Escalation routing, then pilot posting and production validation
Weeks 5–6
Nothing else is required
Deployment, documentation and Agent Care handover are ours.
Delivery timeline
Four phases across six weeks
Phases are drawn over the weeks they actually occupy. Week 5 runs the duration basis against live postings while the evaluation is still open.
Phase
W1
W2
W3
W4
W5
W6
- DiscoveryW1
- BuildW2 – W3
- EvaluateW4 – W5
- Pilot & LaunchW5 – W6
Week focus
W1Block grid, release clock and add-on rules read into the model
W2Scheduling and staffing integration, then the duration basis
W3Constraint checks, preference cards and requirement lists
W4Evaluation suite, parity slices and escalation routing
W5Scheduling write-back, supervised posting and targeted corrections
W6A full block cycle posted beside the existing process, then handover
- Reading the band
- The duration basis is built in week 2 but is not trusted until week 5 measures it against real days, which is why Evaluate and Pilot overlap.
- At the end of W6
- Posting has run alongside the existing process for a full block cycle and the day-of numbers have been read by service line, then Agent Care takes over monitoring.
- Duration
- Six-week plan shown · typical delivery 4–6 weeks depending on scope confirmed in discovery.
Next step · Healthcare AI agent
Build a surgical-scheduling agent around your block rules.
Show us your block grid, your release and add-on rules, and what each case needs before it can go ahead. We'll rebuild your duration estimates from your own case history and replay a past block week.
- Typical delivery
- 4–6 weeks
- Deployment
- Your OR scheduling stack
- Case priority
- Surgeon and anaesthesia
- After launch
- Agent Care available
Nestack Agents · Surgical scheduling & OR utilisation
AGT-HC-09 · Agent Care available after launch