Perioperative operations and value

01 / 19

KairosWorking name

Run the OR better. Manage the whole episode. Prove the savings.

Anesthesia-led perioperative performance: one platform for the operating rooms, every step of the surgical episode, and a value program whose savings a hospital CFO can check line by line.

Download PDF Open the demo

Discussion draft · September 2026 · Prepared by the founding team with the clinical partner · Synthetic data only

The Kairos Overview for Sierra Pacific Medical Center, a fictional hospital: key-indicator tiles for prime-time utilization 62.4%, block utilization 66.7%, first-case on-time starts 64.4%, mean turnover 36.3 minutes and same-day cancellations 3.1%, each with a sparkline, a change versus the prior 30 days and a Show basis link; below, cards for Run the OR better, Manage the episode and Today at a glance. — The Overview, signed in as a perioperative director. Every tile opens the arithmetic behind it.

Why this matters

02 / 19

Surgical care is where cost, risk and capacity meet

The perioperative pathway touches almost every hospital cost line — OR minutes, beds, complications, readmissions — yet it is managed one slot at a time.

~1/3

of U.S. healthcare spending is associated with surgical care

Per the clinical partner’s framework

28–32%

of the global burden of disease is attributable to conditions requiring surgical care

Per the clinical partner’s framework

Risk is discovered too late
Anemia, diabetes, sleep apnea and anticoagulation are too often found on the day of surgery — when the only options are delay, cancel or proceed with more risk.
Incentives are fragmented
Surgeons, anesthesia, the hospital and the payer each optimize their own piece. Nobody is paid to own the whole episode, from the decision to operate to the discharge.
Savings are claimed, not proven
Tools report recovered minutes or projected dollars. A CFO, a surgeon and a payer rarely see the same arithmetic — so shared savings stall at validation.

“I need an AI software that cuts down the cost all around and shows the savings.”
The clinical partner — an anesthesiologist who owns hospitals and clinics in California

Surgical-care share of spending and of the global burden of disease: as cited in the clinical partner's framework (Farm to OR Table deck, 2026). Quote: the clinical partner, September 2026.

The incentive gap

03 / 19

Hourly staffing pays for presence, not pathway performance

The team best placed to change the episode — anesthesia, which sees every patient before, during and after surgery — is paid the same whether the pathway works or not.

1
Today: hourly staffing
Pays for hours covered — the same whether a case was cancelled, ran late or went home early.
2
Base salary
Protects safety and recruitment.
3
Value incentive
Target 15–20% of base, capped, paid on the balanced scorecard.
4
Share of validated savings
Only after clinical and finance validation. Both value components are zero when the safety gate fails.

The proposed model

1
Base
A competitive base protects safety and recruitment. It never depends on volume.
2
Quality and access incentive
A defined upside — target 15–20% of base, capped — paid on a five-domain, risk-adjusted scorecard. No single-metric payment.
3
Validated savings share
A defined share of savings that passed clinical review, then finance review. Nothing is shared before validation.

Safety gate. Any period that misses a minimum safety threshold pays no value component at all — enforced by the software, not a policy PDF.

Compensation model: docs/PRODUCT-VISION.md and docs/VALUE-ENGINES.md §7. The demo's seeded plan (anesthesiologist base $480k, target 17.5%, cap 25%; 30% of validated savings shared) is illustrative.

Farm to OR Table

04 / 19

One accountable team follows the patient through five stages

The clinical partner’s model, made operational: every case sits on a pathway from the decision to operate through discharge, and every stage feeds the scorecard.

1Plan
Risk screen at scheduling
Kairos does
- Risk tier with its basis at the moment of booking
- Duration prediction P50 / P75 / P90
- Block, open-time and slot finder
2Prepare
Optimize, test by protocol
Kairos does
- Open optimization items: anemia, diabetes, smoking, frailty, OSA, anticoagulation
- Each test and consult flagged indicated or not indicated per protocol
3Deliver
ERAS, safe anesthesia, reliable starts
Kairos does
- Live command center: first-case starts, turnover, delays
- Anesthesia coverage and 1:4 medical direction in 15-minute slices
- ERAS elements tracked
4Recover
Analgesia, prevent complications
Kairos does
- DVT/PE, reintubation, kidney injury, severe PONV
- Opioid exposure (MME) per inpatient case
- Risk-adjusted length of stay
5Transition
Discharge planned before surgery
Kairos does
- Discharge destination set pre-op
- 30-day readmission and ED-visit windows
- Episode closes into the scorecard

Measured end to endBalanced scorecard with safety gateSavings ledger with the arithmetic on every lineCompensation statements for the team

Stages: the clinical partner's Farm to OR Table model (docs/PRODUCT-VISION.md). What Kairos does at each stage: docs/VALUE-ENGINES.md §10 and docs/OPERATIONS-MODULES.md.

Positioning

05 / 19

Why this is bigger than block scheduling

Block-scheduling tools optimize one lever and sell it to the periop director. Kairos treats OR utilization as one of five levers — and still does that lever well, deterministically and with its basis shown.

Unit of optimization
Block-scheduling tools: OR block minutes
Kairos: The surgical episode, from the decision to operate to recovery
Buyer
Block-scheduling tools: Perioperative director
Kairos: Health system and anesthesia partnership, under a value contract
Economic story
Block-scheduling tools: Recovered OR capacity
Kairos: Validated episode savings, shared with the clinical team
Clinical scope
Block-scheduling tools: Scheduling
Kairos: Risk screening, optimization, testing stewardship, ERAS, recovery and discharge
Incentive alignment
Block-scheduling tools: None
Kairos: Scorecard-driven compensation: base, bonus and a share of validated savings
AI
Block-scheduling tools: Predictive scheduling
Kairos: Predictive scheduling and a copilot over the whole episode — briefs, recommendations, savings narratives

Block-scheduling tools compared with Kairos
Dimension	Block-scheduling tools	Kairos
Unit of optimization	OR block minutes	The surgical episode, from the decision to operate to recovery
Buyer	Perioperative director	Health system and anesthesia partnership, under a value contract
Economic story	Recovered OR capacity	Validated episode savings, shared with the clinical team
Clinical scope	Scheduling	Risk screening, optimization, testing stewardship, ERAS, recovery and discharge
Incentive alignment	None	Scorecard-driven compensation: base, bonus and a share of validated savings
AI	Predictive scheduling	Predictive scheduling and a copilot over the whole episode — briefs, recommendations, savings narratives

The buyer is the partnership that can change the episode, not only the office that allocates blocks.
The economic story ends in validated dollars with a basis — not a projection.
Multi-facility, multi-organization tenancy from day one: the second hospital is configuration.

Comparison: docs/PRODUCT-VISION.md, generalized to the block-scheduling tool category.

The product · lever 1 of 5 · Run the OR better

06 / 19

Release the time owners will not use, and book cases for how long they really take

Blocks, Release recommendations tab at Valley Community Hospital: 49 blocks recommended for release in the next 28 days, 133 hours recoverable, 107 blocks evaluated; the top recommendation is a surgeon block in OR 5 with 390 recoverable minutes, expected utilization 28.7% against a 75% target, 90 of 600 minutes booked, 8 days of lead time and high confidence 92%, with its basis. — **49 blocks, 7,980 minutes** recommended for release at the pilot hospital over the next 28 days — a deterministic rule with its basis; the owner decides.

Schedule day view at Sierra Pacific Medical Center: a room timeline from 06:00 to 20:00 with twelve operating rooms, block bands, 42 cases as bars coloured by status (done, in progress), a prime-time band and a red now line at 22:18. — The day view: 42 cases in 12 rooms, block bands, prime time and a live now line. Booking shows P50/P75/P90 predictions and the database itself rejects a double booking.

Block utilization per owner, raw and adjusted, with a weekday × hour heatmap.
Release recommendations ranked by recoverable minutes, from the owner’s trailing utilization and the facility’s booking curve.
An open-time exchange, and duration predictions that beat surgeons’ own estimates.

Synthetic data · fictional health system

Blocks → Release recommendations, Valley Community Hospital, next 28 days as of Sep 24, 2026. Schedule: Sierra Pacific Medical Center day view, Sep 24, 2026. Demo script minutes 2–4.

The product · lever 1 of 5 · Run the OR better

07 / 19

Run the day from one screen — including anesthesia coverage

Command center at Valley Community Hospital, live and refreshing every 30 seconds: first-case on-time starts 66.7%, 19 cases done, 1 in progress, 2 cancellations, mean turnover 28.8 minutes, 254 delay minutes today; delay minutes by category led by prior case overrun at 181 minutes; room cards for OR 1 to OR 3 with the current case, projected end against prime time and the anesthesia provider. — The live board: every room’s current case, projected end against prime time, delays by category, turnover and first-case starts.

Anesthesia staffing, Coverage tab: peak 6 rooms running, highest ratio 1:3 against a 1:4 maximum, 0 coverage violations, 15-minute slices; a chart across the day of rooms running versus directing anesthesiologists, with the most rooms any one anesthesiologist covers per slice, and a Violations card reporting none on this day. — Coverage in 15-minute slices: rooms running against directing anesthesiologists, the 1:4 medical-direction ratio, and every flagged interval with its basis.

Record milestones as they happen; delays carry a reason.
Anesthesia staffing is a first-class module, not an afterthought.
The same numbers roll into the reliability domain of the scorecard.

Synthetic data · fictional health system

Command center and Anesthesia → Coverage at Valley Community Hospital, Sep 24, 2026 (signed in as the anesthesia medical director). Demo script minutes 4–6.

The product · lever 2 of 5 · Manage the whole episode

08 / 19

The team manages the episode, not just the slot

Every upcoming case sits on a PLAN → TRANSITION board with its risk tier, open optimization items, testing-stewardship flags and discharge plan — weeks before the day of surgery, when there is still time to act.

Risk screen at scheduling: a score with its basis, never a clinical conclusion.
Tests and consults marked indicated or not indicated per protocol; only non-indicated ones count as avoidable.
ERAS elements, complications and the discharge destination tracked per episode.
Pathway versus legacy care side by side, so the before/after is visible per service line.

Software organizes, measures and drafts; clinicians decide.

Synthetic data · fictional health system

Pathway board at Valley Community Hospital with five columns — Plan (28 cases, risk screen at scheduling), Prepare (129, optimization and evidence-based testing), Deliver (1), Recover (17) and Transition — each card showing a de-identified patient reference, the procedure, the surgeon, the date, a risk tier such as Unscreened, Moderate or High, and flags such as Risk screen, Low-value tests, Optimization, Discharge plan, Pathway or Legacy. — Farm to OR Table as a working board: stage, risk tier and open items on every card.

Pathway board at Valley Community Hospital, surgeries Sep 10 – Oct 15, 2026. Stewardship rule: docs/VALUE-ENGINES.md §9–10. Demo script minutes 6–7.

The product · lever 3 of 5 · Prove the value

09 / 19

Scorecard with a safety gate → savings with their basis → what the team earns

Value program scorecard for August 2026 at the Valley Community Orthopedics pilot: composite 1.05 with 11 of 14 metrics at target, incentive earned 18.31% of base against a 17.5% target and 25% cap, 134 episodes scored and 7 excluded; a green Safety gate passed panel; five domain gauges — Access 0.72, Reliability 1.12, Recovery 0.98, Safety 1.18, Stewardship 1.11; and the metrics table with observed, baseline, target and stretch values. — **Composite 1.05**, gate passed, 11 of 14 metrics at target — five risk-adjusted domains, no single-metric payment.

Savings ledger for the pilot, March 1 to September 23, 2026: gross savings $1.1M as a draft estimate, $0 in clinical review, $0 in finance review, $0 validated and a $0 shared-savings pool; a waterfall of gross savings by category — length-of-stay days $353.4K, complications $237.5K, low-value tests $142.3K, SNF days $137.1K, same-day cancellations $87K, readmissions $75K and smaller lines — and the ledger's review queue with 84 draft lines. — **$1.09M gross, $0 validated** — every line shows baseline rate × volume × unit cost, and nothing counts until clinical and finance review.

Compensation for August 2026: composite 1.05, safety gate passed, incentive earned 18.31%, validated savings $0, base plus incentive $354.4K on a base of $300.3K; statements per anesthesiologist and CRNA with FTE, base, incentive, savings share and total, and a stacked chart by provider type. — Statements per provider: base + incentive + savings share, with FTE, the gate and the arithmetic one click away.

Safety gate fails closed: no eligible episodes means no value payment.
Validation needs two different people: clinical review, then finance review.
The program cannot activate until counsel approves the gainsharing design.

Synthetic data · fictional health system

Value program → Scorecard, Aug 2026, Valley Community Orthopedics pilot (synthetic); Value program → Savings ledger, Mar 1 – Sep 23, 2026, all lines in draft; Value program → Compensation, Aug 2026, Valley Community anesthesia team. Demo script minutes 7–8.5.

The product · lever 4 of 5 · Claude across all of it

10 / 19

Claude over the live data — with guardrails you can see

Copilot conversation scoped to Valley Community Hospital. Question: Which blocks at Valley Community should we release next week and why? The answer, labelled Generated by Claude, an estimate you check, says four block instances fall inside next week worth 720 recoverable minutes and shows a table of date, room, owner, booked and retained minutes, expected utilization and release minutes. — Asked in plain English, answered from read-only tools over de-identified aggregates — violet marks machine-generated content.

The end of the same answer: the method (expected utilization against the booking curve), a note that these are estimates to verify with the owners and that Kairos releases nothing automatically, and an expanded tool trace showing one call to get_release_recommendations with its arguments, 341 milliseconds and 7 rows, plus the model and token counts. — The tool trace under every answer: which tool ran, with what arguments, how long it took and how many rows it returned.

De-identified only

A structural type and a guard on every outbound call block identifiers; a blocked call is logged.

Deterministic numbers

Predictions, recommendations, scorecards and savings are computed by engines; Claude explains, never computes a dollar.

Nothing auto-applied

Releasing a block, booking a case or validating a saving is always a person’s action.

Governed

Model inventory, intended use and human oversight per ISO/IEC 42001 and the NIST AI RMF.

Copilot, signed in as the anesthesia medical director, Valley Community Hospital, Sep 24, 2026; one question asked for this deck. AI governance: docs/AI-COPILOT.md and docs/COMPLIANCE-REGISTER.md (ISO/IEC 42001, NIST AI RMF). Demo script minutes 8.5–9.5.

The product · lever 5 of 5 · The training data underneath

11 / 19

A health system of synthetic data, rebuilt from one seed

45,271

surgical cases

18 months of history + six weeks booked ahead

837,937

rows across 48 tables

CSV or JSON export with a data dictionary

3 · 24

facilities · operating rooms

Two hospitals and an ASC, fictional

69 · 45

surgeons · anesthesia providers

Each with speed, estimate bias, persona

64 s

to regenerate everything

Seed 20260924, generator gen-1.1.0

45.7%

lower error than surgeon estimates

Best facility; all three clear the 30% target

Analytics, Prediction accuracy tab for Foothill Surgery Center over 12 months: Kairos is 45.7% closer to actual in-room time than the surgeon's estimate on the same 9,435 completed cases; mean absolute error 13.6 falls to 7.4 minutes, MAPE 24.9% to 12.9%, within ±15% rises from 34% to 65%, mean bias −9.2 versus −1.2 minutes; a Meets the 30% target badge and a paired bar chart of error by service line. — Kairos P50 versus the surgeon’s own estimate on the same completed cases, with the evaluation target shown.

The Training hub: every module has an interactive tour on the user's own screens, a narrated video recorded from the real app on synthetic data, and a printable quick start; cards for Overview, Blocks, Schedule and Command center with video players, tour length and steps, and Start tour buttons. — Every module ships with an interactive tour, a narrated video recorded from the real app, and a printable quick start.

Training data → summary (gen-1.1.0, seed 20260924, generated Sep 25, 2026); Analytics → Prediction accuracy, trailing 12 months (Sep 24, 2025 – Sep 23, 2026), mean absolute error of Kairos P50 vs the surgeon's estimate on the same completed cases. Nothing is derived from real records (docs/TRAINING-DATA.md).

Proof from the synthetic pilot

12 / 19

Measured, gated, priced and explained — end to end

Valley Community Hospital (fictional) put orthopedics on the pathway on March 1, 2026. Everything below is computed by the platform’s engines from the synthetic dataset.

1.05

August 2026 composite score

Safety gate passed · 11 of 14 metrics at target · 134 episodes scored

18.31%

incentive earned, of base

Target 17.5%, cap 25% — 0 had the gate failed

45.7%

closer than surgeon estimates

42.4% and 31.5% at the other two facilities; target ≥ 30%

7,980 min

recommended for release, next 28 days

49 of 107 blocks at the pilot hospital

Five domains, August 2026 — risk-adjusted; the tick is target (1.00)

AccessProgressing
ReliabilityAt or above target
RecoveryProgressing
SafetyAt or above target
StewardshipAt or above target

$1.09M gross savings identified Mar–Sep across 10 categories and 84 ledger lines — led by length-of-stay days avoided ($353K), complications avoided ($237K), low-value tests avoided ($142K). $0 validated so far: no line counts until a clinician and then finance sign it off.

Baseline year → August 2026
Metric	Baseline	Aug 2026
First-case on-time starts	65.5%	85.3%
Low-value pre-op tests	36.7%	11.8%
Risk-adjusted LOS (O/E)	1.01	0.87
Opioid MME per inpatient case	261	155
30-day readmission	7.4%	4.2%

Synthetic, and honest about it. The generator builds a pathway effect into the pilot service line. These figures show that the engines measure it, gate it, price it and explain it correctly from end to end — they are not evidence that the pathway works. The pilot is where that evidence comes from.

Sources: Value program → Scorecard, Aug 2026, Valley Community Orthopedics pilot (synthetic). Analytics → Prediction accuracy, trailing 12 months (Sep 24, 2025 – Sep 23, 2026), mean absolute error of Kairos P50 vs the surgeon's estimate on the same completed cases. Blocks → Release recommendations, Valley Community Hospital, next 28 days as of Sep 24, 2026. Value program → Savings ledger, Mar 1 – Sep 23, 2026, all lines in draft. Metric pairs: August 2026 observed vs the program’s measured 12-month baseline (Mar 2025 – Feb 2026).

Built for hospitals from day one

13 / 19

Built to hospital standards — evidenced, not yet certified

The Compliance page's live control register: 39 controls, 29 implemented, 6 infrastructure, 4 at pilot, 0 planned, 28 evidence checks passing and 1 failing; last evidence run Sep 24, 2026; framework coverage bars for ISO 27001, SOC 2, NIST CSF 2.0, ISO 27799, NIST AI RMF, Cures Act, CMIA, WCAG 2.1 AA, HIPAA, HITRUST CSF, ISO 42001, HHS 405(d), California H&SC §1280.15, FDA CDS and AKS/Stark. — **39 controls**, 28 runtime evidence checks passing and 1 failing — shown, not hidden. Compliance is a live system, not a PDF.

HIPAA by design

Identifiers live only in an encrypted vault (AES-256-GCM); one service decrypts, and every decrypt is logged.
PostgreSQL row-level security: a query without a tenant returns no rows.
Append-only, hash-chained audit; the database refuses updates and deletes on it.
15-minute automatic logoff, lockout, break-glass access with review.

HIPAA and HITECH
Privacy, Security and Breach Notification Rules · implemented per NIST SP 800-66 r2
Designed & evidenced
HITRUST CSF v11
e1 → i1 → r2
At pilot
SOC 2 Type II
Security · Availability · Confidentiality · Privacy
Designed & evidenced
ISO/IEC 27001, 27002 and 27799
ISMS · Annex A · health informatics
Designed & evidenced
ISO/IEC 42001 and NIST AI RMF
AI management system · AI risk
Designed & evidenced
NIST CSF 2.0 and HHS 405(d)
Govern · Identify · Protect · Detect · Respond · Recover
Infrastructure
California CMIA, H&SC §1280.15 and CCPA/CPRA
Civil Code §56 · Health & Safety Code §1280.15 · Civil Code §1798.82
At pilot
21st Century Cures Act
ONC information blocking
Designed & evidenced
FDA Clinical Decision Support guidance
Non-device CDS (2022)
Designed & evidenced
Anti-Kickback Statute, Stark Law and CMP gainsharing
OIG gainsharing safeguards
Designed & evidenced

Targets: HITRUST e1 readiness at pilot, i1 in year one; first SOC 2 Type II window after the pilot. WCAG 2.1 AA on every page.

Standards and status: docs/COMPLIANCE-REGISTER.md §1–3 and the public /compliance page. Control counts: Compliance page (admin), last evidence run Sep 24, 2026. No certification or audit has been performed.

Architecture at a glance

14 / 19

One API owns every rule; the AI sees de-identified aggregates only

1
Web — Next.js
A pure client of the API: no database, no secrets, httpOnly session cookie.
2
API — Spring Boot
Owns tenancy, roles, audit, encryption, the engines and every AI call; FHIR R4 façade.
3
PostgreSQL 16 with row-level security
No tenant set means no rows; identifiers encrypted in a vault; append-only, hash-chained audit.
4
Claude
De-identified aggregates only, checked on every call; tool traces stored; nothing auto-applied.
5
EHR connectors
FHIR R4, SMART Backend Services, Bulk Data and HL7 v2 SIU/ADT.

FHIR R4 + US Core

15 resource types served for read and search; everything exportable as FHIR, NDJSON or CSV.

HL7 v2 SIU / ADT

Scheduling feeds through the hospital’s integration engine create, update and cancel cases.

SMART Backend Services

System-to-system access to the EHR; vendor connectors await registration at the pilot site.

Infrastructure that scales

Containers across two availability zones, managed encrypted PostgreSQL, cross-region snapshots, keys in a cloud KMS.

Architecture: docs/ARCHITECTURE.md. Interoperability: docs/INTEROPERABILITY.md (15 FHIR R4 resource types in the live CapabilityStatement). Infrastructure: docs/DEPLOYMENT.md; the 99.9% availability figure is a design target, not a measured service level.

Commercial design

15 / 19

Gainsharing with safeguards — and a revenue model to agree together

The value program follows the clinical partner’s five-step gainsharing design; the software enforces the OIG safeguards so counsel reviews a method, not a promise.

1
Baseline
12 months per service line
2
Safeguards
Risk adjustment, exclusions, safety gate
3
Savings
Against an agreed counterfactual
4
Validated share
Only after clinical and finance review
5
Reinvest
Pathway staff, analytics, clinician incentives

Safety gate
Risk adjustment
No single-metric payment
Indicated-only stewardship
No payment without validation
Caps and a documented method

How Kairos earns — options for discussion Not decided

Option A · Platform subscription
An annual subscription per facility (tiered by operating rooms), covering the operations modules, the pathway, the value program, the copilot and support.
Option B · Subscription + success component
A lower subscription plus a defined, capped component tied to independently validated savings — only after the hospital’s own clinical and finance review.
Option C · Implementation
A one-time fee for the EHR connection, the 12-month baseline load and value-program set-up, alongside either option above.

Gainsharing steps: docs/PRODUCT-VISION.md. Safeguards as implemented: docs/VALUE-ENGINES.md §9. Revenue options are for discussion only — nothing is decided, and any success-based fee needs the same counsel review as the gainsharing design.

Pilot plan

16 / 19

One hospital, one service line, twelve months

Start at one small hospital with a high-volume elective service line, measurable variation and engaged surgeons. Multi-facility tenancy exists from day one, so the second hospital is configuration, not engineering.

Pilot phases, the clinical partner’s plan and what Kairos delivers
Phase	The clinical partner’s plan	Kairos delivers
0–90 days	Choose one service line; baseline metrics; map the workflow; define governance	Connect the OR schedule (HL7 v2 or FHIR); load 12 months of history; agree metric definitions, exclusions and safety thresholds; data-use agreement and legal review of the gainsharing design
Months 4–6	Build the pathway: pre-op triage, ERAS order sets, discharge planning	Pathway board live for the service line; risk screen at scheduling; testing-stewardship flags; release recommendations and open-time exchange; command center and anesthesia coverage
Months 7–12	Measure and refine: monthly dashboard, case review, validate savings	Monthly scorecard with the safety gate; savings through clinical and finance validation; compensation statements; CFO narrative; prediction-accuracy review
Year 2	Scale what works: add procedures, expand partners, introduce the incentive pool	Second service line and second hospital as configuration; multi-facility benchmarks; HITRUST i1 and SOC 2 Type II windows; incentive pool activated

What the pilot needs from the hospital

An HL7 v2 SIU feed or FHIR R4 access to the OR schedule
12 months of history for the baseline
Agreed metric definitions, exclusions and safety thresholds
A data-use agreement and business associate agreement
Counsel’s review of the gainsharing design
A clinical reviewer and a finance reviewer for validation

Phases: the clinical partner's twelve-month pilot, mapped to Kairos deliverables in docs/PRODUCT-VISION.md. Hospital requirements: docs/DEMO-SCRIPT.md talking points.

The partnership

17 / 19

Clinical credibility and the ability to build and run it

The founding team

Large-scale system build-out and healthcare experience

Builds and runs the platform: product, engineering, data and AI
The security, privacy and compliance program and its evidence
Cloud infrastructure, integration with the hospital’s EHR
Commercial operations and the path to certification

The clinical partner

An anesthesiologist who owns hospitals and clinics in California

OR-optimization expertise and the Farm to OR Table model
The pilot site, the service line and clinical governance
Metric definitions, protocols and case review with the team
Relationships into large hospital systems for year two

Together

Design the pilot and the value program, sit on its governance, and take a proven result to the large systems the clinical partner already knows.

To be agreed

Equity, funding, roles and governance of the company.

Roles as discussed; nothing here is a commitment. Names are intentionally omitted.

What exists today

18 / 19

A working platform, built in one intensive build cycle

14

modules, end to end

API, web, tests and docs for each

39

compliance controls

with runtime evidence checks

15

FHIR R4 resource types

plus HL7 v2 SIU/ADT ingest

13

narrated training videos

and 11 interactive tours

45,271

synthetic cases

regenerated from a seed in about a minute

Built and demonstrable

Overview
Blocks
Schedule
Command center
Anesthesia staffing
Pathway
Value program
Analytics
Copilot
Interoperability
Training data
Compliance
Training
Settings

Seven demo roles, three fictional facilities, multi-organization tenancy, a public site with security, privacy, terms and compliance disclosures.

Second iteration — before real patient data

Live EHR connectors at the pilot site (vendor registration, integration-engine feed)
Multi-factor sign-in required for anyone with patient-data access
Envelope encryption with managed keys for the identifier vault
A dedicated cloud account with full account-level logging and threat detection
Business associate agreement and zero-retention terms with the AI vendor
A named privacy officer and a tested incident exercise
Demo accounts removed; data-use agreement and a data-quality gate on real feeds

The Training hub with a card per module — Overview, Blocks, Schedule, Command center — each offering a narrated video recorded from the real app on synthetic data, an interactive tour with its length and steps, and a printable quick start. — Every module ships with a tour, a narrated video and a quick start.

Module list: the signed-in navigation. Counts: Compliance page, FHIR CapabilityStatement, training manifest, Training data summary. Before-real-data list: docs/RISK-REGISTER.md (actions due before the pilot) and docs/INTEROPERABILITY.md.

The ask · next steps

19 / 19

Five decisions to start the pilot

1
Agree the pilot site and service line
One hospital; one high-volume elective line with measurable variation and engaged surgeons.
2
Sign a data-use agreement
With a business associate agreement, before any real data moves.
3
Counsel reviews the gainsharing design
Anti-Kickback, Stark and CMP review of the value program — required before it can activate.
4
Open EHR access
An HL7 v2 SIU feed or FHIR R4 access, and 12 months of history for the baseline.
5
Decide the brand name
Kairos is a working name.

Kairos

Software organizes, measures and drafts; clinicians decide. Let’s prove it at one hospital.

Contact: the founding team

Download PDF Open the demo

Kairos is a working name for software in beta. Every product figure in this deck comes from a synthetic, fictional health system; none is a result at a real hospital. Market figures are as cited in the clinical partner’s framework. Kairos is not a medical device and does not provide medical, legal or financial advice. This deck is a discussion draft, not an offer of securities.