Recover
Under-coded claims surfaced with the correct value and the evidence: money you'd already earned, kept.
See itcoming.
xuman AI™ builds foresight for Indian hospitals: xuman catches the claim error before it’s submitted, and Aurora, in development, is being built to forecast the ICU crisis hours ahead. Deployed inside your hospital, never outside it.
2,906 FILES · ONE-MONTH HOSPITAL PILOT · UNDER 1 HR TO FLAG
xuman consoleclaims desk
ON-PREMDPDP
FILEPACKAGEBOOKEDINDICATEDVERDICT
XU-••38IPD · medical₹••,•••—CLEAN
XU-••39Day-care₹•,•••—CLEAN
XU-••41IPD · surgical₹••,•••—QUERY
XU-••42IPD · surgical · Gr II₹5,000₹12,000SCANFLAGFIXED
⚑ PACKAGE UNDER-VALUED · booked ₹5,000 against an indicated ₹12,000
✓ APPLY ₹12,000 PRE-SUBMISSION+₹7,000
AUDIT sha256 ⋯8f2c1d · CHAIN ✓ · DE-IDENTIFIED · CAUGHT 5× IN APRIL
XU-••43IPD · medical₹••,•••—CLEAN
2,906 FILES · ONE-MONTH PILOTILLUSTRATIVE · FROM A REAL CATCH
PILOT RECORD
One-month pilot inside a working Indian hospital. The partner stays unnamed by design.
2,906
claim files processed
₹5,000 → ₹12,000
one under-valued package, caught 5× in April
<1 hr
from file received to error flagged
DPDP
ready by design, hash-chained audit
Healthcare’s costliest failuresaren’t failures of skill.They’re failures of foresight.
xuman
IN PILOTCLAIMS INTELLIGENCE
The claim error caught before it’s submitted: coded right, documented complete, approvable the first time.
SEE THE CATCH
Aurora
IN DEVELOPMENTICU INTELLIGENCE · PRE-VALIDATION
The crash has a signature. Aurora is being built to read it hours early, with the decision always in the clinician’s hands.
SEE THE FORECAST
xumanCLAIMS INTELLIGENCE
IN PILOT · REAL FIGURES ONLYMoney you’ve already earnedshouldn’t die in a query.
Recovery teams start working after the rejection arrives. xuman starts before submission: the extraction layer reads every line of the file, prices the package right, and completes the record while it can still be fixed.
Close
The documentation gaps that trigger queries, found and closed while the file is still on your desk.
Predict
Every claim scored CLEAN, QUERY, or REJECTION: the outcome known before you send.
THE VALIDATION PIPELINE
Six stages between a file and its verdict.
01READ
The extraction layer reads every page.
02RECONCILE
Every source agrees before it moves.
03PRICE + MATCHDPME
The package priced right.
04VALIDATE
Complete before it leaves.
05ADVISE
The fix, not just the flag.
06PREDICT
The outcome, before you send.
PREDICTED PATH →CLEANQUERYREJECTION
See a claim caught live, on your own files.
Book a walkthrough
IN DEVELOPMENT · PRE-VALIDATION
The crash has a signature.Aurora reads it early.
Forecast, not hindsight. Aurora is being built to read the ICU’s earliest signals and hand back the one thing no unit has enough of: time.
THE DECISION STAYS WITH THE CLINICIAN
DEVELOPMENT TARGETS — NOT VALIDATED RESULTS
up to 48 hrs
AKI — DESIGNED FORECAST WINDOW
6–12 hrs
SEPSIS — DESIGNED FORECAST WINDOW
Calibrated
Risk you can trust at 3 a.m.: probabilities that mean what they say.
Explainable
Every forecast shows its drivers. An advisor, never a black box.
Self-hosted
On your own servers, inside your own walls. Data stays with the hospital.
RISK FORECAST
ILLUSTRATIVE — NOT CLINICAL DATA
The published evidence Aurora builds on.
INDEPENDENT RESEARCH — NOT AURORA’S RESULTS- ~0.82 AUROC1
- machine-learning sepsis prediction, against NEWS2 at 0.72 and qSOFA at 0.61
- 0.80–0.90 AUROC2
- gradient-boosting models for acute kidney injury
- 90.2%3
- of dialysis-requiring AKI predicted 48 hours ahead
- 1. Yadgarov et al., 2024 — meta-analysis, ML sepsis prediction.
- 2. Published literature — gradient-boosted AKI models.
- 3. Tomašev et al., DeepMind, Nature 2019.
Aurora is in development and pre-validation. These figures describe published research on comparable approaches; they are not Aurora’s performance, and Aurora is not a certified medical device.
Follow Aurora from first principles to the bedside.
Book a walkthroughYour data never leavesyour hospital.
Both products run where the data lives. Files are de-identified before analysis, every action is written to a tamper-evident audit chain, and the perimeter of the system is the perimeter of your building.
HIS FILES
claims, notes, vitals
DE-IDENTIFY
identifiers masked
ANALYSIS
the extraction layer
VERDICT
flag, fix, audit
NOTHING CROSSES THE WALL · DATA STAYS WITH THE HOSPITAL
- DPDP by design
- Built for India's Digital Personal Data Protection Act from day one, not retrofitted for it.
- Self-hosted
- Deploys on your servers, inside your own walls. Data stays with the hospital.
- De-identified first
- Patient identifiers are masked before any analysis touches the file.
- Hash-chained audit
- Every read and every fix lands on a tamper-evident chain, with erasure built in.
Built inside the hospital,not the boardroom.
xuman AI is founded by a clinician who trained on these wards: Pharm.D at NIMS University, clinical intern at NIMS Hospital, and a member of the core organising team of the WHS Regional Meeting 2025. The pharmacy AI and the TPA claim platform behind xuman weren’t imagined for hospitals; they were built inside one, for the people already working there.
- Clinician-built
- Designed by a clinician who lives these workflows, not imagined from outside them.
- Already shipped
- An AI pharmacy system and a TPA claim platform, built and running inside a working hospital.
- India-first
- DPDP, TPA workflows, and Indian package rates are the starting point, not an afterthought.
See it coming.
Two products. One instinct. See them working on your own files, on a call, on your terms.
- DPDP-READY
- SELF-HOSTED
- HASH-CHAINED AUDIT
- CLINICIAN-BUILT