01 — Health & Life Sciences
From molecule to member: a value engine that models trial economics, care pathways, patient access and commercial performance as one connected system — under GxP, HIPAA and EU AI Act discipline.
North Star
Cost per validated patient outcome
The value engine, end to end
The same six moves we run everywhere, expressed in the physics of this industry.
Where does clinical and commercial value actually pool?
Value pool map by therapy area
What causes cycle time, adherence and access to move?
Causal model of trial + pathway
What happens if we change site mix or protocol?
Trial + care pathway twin
Who executes the change every day?
Multi-agent operating layer
Is it validated, auditable and safe?
GxP + EU AI Act control file
Does the gain hold next quarter?
Outcome contract + drift monitor
Where the value is
Before a twin is built, we agree where the money sits and what moves it.
Value pool
Clinical development
Driver
Trial cycle time and site productivity
Measured as
Days saved × cost of delay
Value pool
Patient access
Driver
Time to therapy, prior-auth throughput
Measured as
Scripts converted per 1,000 referrals
Value pool
Care delivery
Driver
Length of stay, readmission, no-shows
Measured as
Bed-days and avoidable episodes
Value pool
Commercial
Driver
Field effort allocation, HCP engagement quality
Measured as
Revenue per rep-hour
Use cases
Each one is a number somebody already owns, a twin that models it, and an agent that moves it inside a defined authority. Filter by function.
14 use cases
Reads referral letters, faxes and scans, extracts the clinical detail, and books the patient into the right pathway without a queue in between.
Forecasts census, length of stay and discharge timing so staffing and beds are planned against next week rather than last week.
Identifies who is likely to return and what to change while the patient is still in the building, where intervention is cheapest.
Watches the pattern across wards, staff movement and procedures, and raises the signal early enough to break the chain.
Prioritises studies most likely to contain an urgent finding so the radiology queue is ordered by risk, not by arrival time. Support, never a substitute for the clinician.
Surfaces guideline-grounded options at the point of decision, every recommendation carrying the source paragraph behind it.
Detects overpayment and abusive billing patterns across providers and members without freezing legitimate claims in the process.
Prices plans against forecast utilisation and cohort risk rather than last year's book, so margin does not depend on the renewal cycle guessing well.
Predicts which balances are collectible and which route collects them, and drafts the patient communication that works.
Ranks sites on realistic enrolment rather than reputation, and rebalances recruitment while the curve can still be changed.
Reads case reports, literature and public channels for emerging safety signals and routes them into the regulated workflow with the evidence intact.
Mines patient and clinician text for effectiveness and tolerability patterns that structured data alone does not carry.
Allocates field and medical effort to the accounts where engagement changes behaviour, inside compliant content boundaries.
Forecasts encounter volumes for seasonal and emergent conditions so capacity, staffing and supply are positioned ahead of the wave.
Outcome twins
Each twin owns a set of levers. Each feature moves one lever. Each agent executes inside a defined authority.
Simulate protocol, site mix and enrolment before a single patient is screened.
Levers
Features
Multi-agent layer
Enrolment Agent
Rebalances recruitment across sites weekly.
Data Query Agent
Drafts and resolves EDC queries with audit trail.
Model the patient journey end to end and find where outcomes and dollars leak.
Levers
Features
Multi-agent layer
Scheduling Agent
Fills capacity and pre-empts no-shows.
Care Gap Agent
Closes overdue interventions with clinician approval.
Turn prior authorization and payer friction into a measured throughput system.
Levers
Features
Multi-agent layer
Prior-Auth Agent
Assembles and submits payer-ready evidence bundles.
Allocate field and medical effort to the accounts that move outcomes.
Levers
Features
Multi-agent layer
Field Planning Agent
Re-plans territory effort against outcome value.
Unit economics
Set your volume, switch features on or off, and move each impact to your own evidence. Everything is expressed per patient episode.
Episodes per year
120,000
Contribution per patient episode
$800
Episode economics; substitute trial patients or scripts as the unit where relevant.
Annual program investment
$3.5m
Clinical Trial Twin
Unit cost — Cost of delay avoided
4.0% — Reallocates recruitment before the enrolment curve flattens.
$16.3m
annual value
Care Pathway Twin
Unit cost — Length of stay
3.5% — Removes waiting time between clinical decision and discharge.
$14.3m
annual value
Care Pathway Twin
Volume — Utilised capacity
3.0% — Fills slots that would otherwise be lost the same day.
$2.9m
annual value
Access Twin
Leakage — Denials and abandonment
25.0% — Reduces value lost between prescription and therapy start.
$13.9m
annual value
Commercial Twin
Unit revenue — Revenue per episode
1.5% — Better mix and adherence on the same volume.
$7.6m
annual value
Value by feature
Annual value contributed by each active feature at current settings.
Value and ROI vs volume
How engineered value and ROI move as episodes per year change. The marker is your current setting.
Cumulative value vs investment
Where the curve crosses the investment line is payback.
Value mix by lever
Which lever the engineered value is actually coming from.
Baseline contribution
$96.0m
Engineered annual value
$54.9m
ROI on program
1469%
Payback
0.8 months
Directional model. In an engagement every number here is replaced by your measured baseline, signed off by finance, and written into the outcome contract before any agent goes live.
Scale
Prove the causal model on a bounded population with a validated baseline.
Reuse the twin, swap the data contracts, keep the control file.
Shared agent runtime, per-market guardrails, one accepted number.
Integrations
Vendor-agnostic by design. Connectors are added per engagement — no platform lock-in.
Clinical
Commercial
Data & cloud
AI
Regulations and standards
Compliance is a design input, not a review gate. Every agent action is logged, attributable and reversible.
HIPAA
PHI handling, minimum necessary, BAAs across every agent hop.
GDPR
Lawful basis, special-category data, DPIA per automated decision.
EU AI Act
High-risk classification for clinical decision support; logging and human oversight.
GxP / 21 CFR Part 11
Validated systems, e-signatures, immutable audit trail.
MDR / IVDR
Boundary control so no agent becomes an unregistered medical device.
Standards supported
Next industry
Financial Services
