Killa Kunan System AI
Let's talk

Killa Kunan System AI (KKS)

AI proposes. Rules verify. A physician signs.

KKS is an Argentine company that designs AI-assisted medical audit for health payers in Latin America. Our work starts where claims and clinical records are scattered across many providers, and where claims are often paid before a physician has read the full clinical record.

Talk to us about a pilot or partnershipSee how we work
  • Cerberus
  • Pulse

AI finds the patterns. People weigh the context, the process and the consequences.

What AI does

  • Gathers and organizes scattered data.
  • Analyzes large volumes of information.
  • Searches for patterns and signals.

What people weigh

  • The context of each case.
  • The process.
  • The people involved.
  • The consequences of each decision.

The Latin American context

Health care in the region is financed by social health insurance funds (in Argentina, obras sociales), private prepaid health plans (prepagas) and public systems. They share three conditions.

  1. Fragmented information

    A single member's claims, contracts and clinical records sit with different providers, in different formats, and nobody holds the complete picture.

  2. Review by hand and by sample

    Audit teams are small and review manually. Many claims are paid before anyone has compared them with the contract and the clinical record.

  3. Decisions without their evidence

    When management, a provider or a court asks why something was paid or denied, gathering the documents behind that decision takes days.

We designed Cerberus and Pulse for this reality, with Spanish-language clinical records and the rules of each country.

Cerberus. Medical audit before payment and before authorization

Cerberus reviews what a payer is about to pay and what it is about to authorize, against the contract, the clinical record and current clinical evidence.

  • Pre-payment review

    Each claim is read in full before the provider is paid: what was agreed, what was billed, and what the clinical record shows was actually delivered.

  • Prior authorization review

    Requests for high-cost care, disability services, complex imaging and biologics are documented with the clinical evidence and the coverage requirements of each case before a decision is made.

Four lenses on every case

  • Administrative and contractual

    What the contract, the agreement and the payment model allow.

  • Clinical evidence

    Medical necessity, read in the clinical documentation and current guidelines.

  • Legal risk

    Legal aspects of each case that affect quality of care, patient safety and the defensibility of the decision.

  • Cost centers

    Costs by type of contract, agreement and payment model.

A physician auditor signs every finding.

The rules come from years of medical audit practice. Cerberus does not replace the audit team: it extends its reach and gives every decision its evidence.

How we work with a payer. A joint process, not a software sale

We start with a retrospective diagnostic on claims the payer has already paid (post-payment review). It does not interrupt current operations or the relationship with providers, and clinical data stays with the institution that holds it.

With that result, management decides with its own evidence how to continue:

  1. We review current processes with management.

  2. Retrospective diagnostic on claims already paid.

  3. We design processes around the payer's specific problem.

  4. We leave a reliable, queryable database for management decisions.

  5. We support the change in day-to-day work.

Talk to us about a pilot

Pulse. Reading the clinical record for clinical research and patient safety

Pulse reads the clinical record: what happened to each patient, in what order and at which provider.

Patient identification for clinical research

Clinical research needs to find the right patients where they are treated. Today that search depends on the memory of clinical teams or on reviewing records by hand.

Pulse applies a protocol's eligibility criteria to the clinical records and finds, inside the institution that holds the data, the patients who could take part in a study. Before committing to a study, the institution knows how many of its own patients meet the criteria.

The decision to include each patient always belongs to the investigator. At most, an aggregate, de-identified report goes to a sponsor.

Quality of care and patient safety

With the same information, Pulse shows how patients are cared for across a provider network: where a condition was noted but never diagnosed, where a diagnosis did not lead to the expected treatment, and where continuity of care was lost.

Data stays with the institution that holds it

Pulse runs where the clinical records are kept. Identifying data does not leave the institution; findings are reviewed by its own professionals.

Pulse does not police, deny or penalize. It answers to a medical value: caring for and protecting the patient.

Fictional case, no real patient data

Three providers, a diagnosis nobody coded

Roberto, a member in his late fifties, is seen by three providers of different levels of care.

  1. March, emergency department at Provider A: he consults for a headache with high blood pressure. The note says «probable hypertension» and he leaves with no diagnosis and no treatment.

  2. June, Provider B: he is admitted for pneumonia. Nurses record high blood pressure every day. He is discharged with a pneumonia diagnosis, no hypertension and no blood pressure medication.

  3. August, laboratory at Provider C: tests show early kidney damage. Nobody refers him.

  4. November, Provider A: he has a stroke.

Each provider saw one part; none saw the whole path. Pulse brings the four moments together so that a professional can review the case in time, before a serious event.

The same query, applied across a network, finds patients in the same situation: to review their care in time, or to invite them to take part in a clinical research study, always with professional review.

Pulse is a prototype with a completed feasibility test. We are looking for partner institutions for a pilot with independent clinical validation.

How we govern AI

Human control is part of the architecture, not a step added at the end.

Every case follows the same path

  1. We gather documents in any format and record where each one comes from.

  2. We extract the data that matters.

  3. We verify it with rules drawn from real audit practice.

  4. A professional reviews, signs and decides.

  • Human in the loop, by design

    AI proposes, rules verify, and a physician signs. No finding reaches a provider or a member without a professional's decision.

  • Traceable findings

    Each finding carries its source document, the rule applied and the clinical evidence behind it, so it can be explained and audited.

  • No rule changes without human approval

    Clinical evidence and regulations are updated continuously, and a person approves every change to the rules.

  • Data protection under each jurisdiction's law

    We process personal data under the law of each country: Argentina's Personal Data Protection Law 25.326 and, in Brazil, the LGPD.

Team

Dr. Daniel Garmendia

Founder

Founder and owner of Killa Kunan System AI (KKS) and its brands. Physician (nephrology and critical care), medical auditor and MSc in health economics, who designs and develops AI solutions applied to health. Former consultant to the Pan American Health Organization (PAHO).

Dr. Blanca Estela Pereyra

Advisor to Cerberus and Pulse

Associate professor of Foundations of Diagnosis at the School of Medicine of a national university in Argentina. Former consultant to the Pan American Health Organization (PAHO).

Work with us

We are looking for partners to take this work further. Tell us which door is yours.

  • Payers and insurers

    A pilot on your own claims, starting with a retrospective diagnostic.

  • Technology and channel partners

    Physician-led audit and clinical record reading for your platform or your clients in Latin America.

  • Clinical research and public health institutions

    A Pulse pilot with independent clinical validation, with data that stays in your institution.

contacto@kks-ia.com