For me or my family
Tutus, in Spanish: your checkup with samples taken at home or at a partner lab, your risk on five fronts and 20 minutes with a Corpus cardiologist. COP 350,000, all included.
The primary care layer for Latinos around the world
Fewer heart attacks. Lower cost. More life.
Corpus AI is the primary care layer for Latinos around the world: it reads the clinical data that already exists, calculates each person's real risk of heart attack, stroke, hypertension, diabetes and chronic kidney disease, and tells the care team whom to see today. Validated in 382,589 Latin American patients and in 626,396 US adults, without retraining.
- Risk calculated for each personHeart attack, stroke, hypertension, diabetes and kidney
- A care plan made for youWhat to look after, and when
- The medical board reviews every resultCorpus physicians, not an algorithm alone
- Support over WhatsAppNo new apps to learn
- Paper · IJCRP 2026
- Validated · NIH All of Us
HIPAA · GDPR · TRIPOD+AI
- 550+acute cardiovascular events prevented
- ↓20%monthly heart attack incidence
- $2.3M USDsaved
- 3.5Mpatients analyzed
A real production deployment with a health-system partner in Latin America. Name withheld for confidentiality.
Two paths
How do you want to start?
For organizations
Insurers, providers and employers prioritize their population by real risk.
In the press
What has been published about Corpus AI.
Each card links to the original piece, in Spanish.
El emprendedor que busca predecir infartos con un algoritmo desarrollado por colombianos
Corpus AI convierte una historia familiar en una apuesta para prevenir infartos
Corpus AI: la startup colombiana que previno 467 infartos con IA
Corpus AI: la compañía colombiana que utiliza inteligencia artificial para salvar vidas, “hemos contenido 467 infartos”
Inteligencia artificial puede prevenir infartos en Colombia: no todo es crisis en salud
Juan Felipe Forero: medicina predictiva e IA en salud- Corpus AI: startup colombiana previene 200 infartos con IA y datos locales
La compañía colombiana que utiliza inteligencia artificial para prevenir infartos, ACV y otras emergencias cardiovasculares
Corpus AI busca prevenir infartos y ACV en la población
Enfermedades crónicas concentran buena parte del gasto en salud y aumentan el reto para los sistemas sanitarios
The platform
One risk engine, three groups of services.
They all run on the same calibrated risk engine, and each organization turns on the ones it needs.
Risk intelligence
Ranks the population by real risk and splits it into cohorts, each with its clinical action, from clinical data that is structured and standardized.
Care and access
Appointment scheduling with reminders, transportation to the appointment and transcription of the visit into a structured note.
Ongoing support
A care plan for each person, habit tracking and risk reassessment with every new data point, over WhatsApp.
How it works
One population, four decisions.
The engine reads a population's clinical data and places each patient in one of four cohorts. Each cohort comes with its action: there is no score to interpret.
1,248,302 patients · 30-day window
Immediate action
Escalate · 24 h
8.2%102,361 patients
High-risk intervention
Contact · 7 d
17.9%223,446 patients
Active chronic management
Care plan
33.6%419,429 patients
Stable monitoring
Passive follow-up
40.3%503,066 patients
Result in production
- ↓20%monthly heart attack incidence
- $2.3M USDsaved
A real production deployment with a health-system partner in Latin America. Name withheld for confidentiality.
Evidence
The heart attack model is published and in plain sight.
The model estimates the risk of a heart attack at 6 and 12 months from routine clinical data. It was derived and validated in an integrated cardiovascular cohort in Colombia; the open-access article is here in full.
Published article · IJCRP 2026
A next-generation prediction risk model for acute myocardial infarction: Derivation and validation in a multi-centre cohort
Amorocho-Morales JD, Parra Guevara S, Quintero-Muñoz E, Dimas G, Correa-Morales JE. Int J Cardiol Cardiovasc Risk Prev. 2026;30:200659. Published online on June 8, 2026.
© 2026 The Author(s). Published by Elsevier B.V. Open access under CC BY 4.0.
DOI 10.1016/j.ijcrp.2026.200659
Figures from the article
- 382,589patients in Colombia
- O/E 0.998calibration, in internal validation
- AUC 0.869discrimination
O/E compares the heart attacks observed with those the model expected: 1 is perfect calibration.
External validation of the same model · not in this article
A separate validation, in another population
The same model, without retraining, was validated with US data. It is an analysis separate from the article: its figures do not come from it.
- 626,396adults
- NIH All of UsUSA
- O/E ≈ 0.98calibration
Data
How we handle health data.
We process personal data under Colombia's Law 1581 of 2012.
- ColombiaLaw 1581 of 2012: Personal data protection
United StatesHealth data under the US standard
European UnionPersonal data protection, EU standard
We comply with HIPAA (United States) and GDPR (European Union) to protect health data.
About us
Who is behind Corpus AI.
Juan Felipe Forero
Chief Executive Officer
Leads the strategy, the relationship with institutional clients (payers, providers, insurers) and the commercial operation of Corpus AI.
Santiago Peláez
Chief Operating Officer
Owns operational execution, the rollout with each client and the internal processes that sustain clinical delivery.
Juan Bernardo Benavides
Tech Lead
Leads platform architecture, EHR integrations and the engineering discipline behind every prediction: idempotency, determinism, auditability.
Dr. Jeffrey Castellanos
Chief Medical Officer
Internist and clinical cardiologist in the heart failure program at Clínica Universitaria Colombia. MBA. Responsible for the model's clinical validity.
José David Amorocho-Morales
Chief AI Officer
PhD(c) in psychology applied to AI. Leads the development of the predictive model, calibration and validation. Lead author of the article published in IJCRP 2026.
“We have been in the cardiovascular fight for 80 years, and it frustrated me to see that we are losing it. That is why Corpus AI exists: we are here to show that a Latino's risk is not the same as that of a 1.90 m white European. A 1.60 m mestiza woman also deserves evidence-based medicine made to her measure. That is why we measure you against our own population.”
Advisory Board
Physicians who review what we do.
The physicians on our Advisory Board are with us throughout the process: they review what we do and watch over the model's clinical validity and its implementation in the care workflow.
Dr. Henry Gallardo
Hospital Management · General Director · Fundación Santa Fe de Bogotá
General Director of Fundación Santa Fe de Bogotá. He brings the perspective of someone who runs a high-complexity hospital: how prediction gets implemented in real institutions without breaking the clinical workflow.
Dr. Augusto Galán Sarmiento
Cardiology · Public Health · Former Minister of Health of Colombia
Cardiologist and former Minister of Health of Colombia. He watches that the model makes sense at health-system scale: populations, insurance, and public prevention policy.
Dr. Guillermo Ortiz
Critical Care · Intensive Care Medicine · Universidad El Bosque · Centro Policlínico del Olaya
Internal-medicine physician, intensivist, and Latin American reference in critical care. His clinical review keeps the model’s cohorts meaningful for patients who pass through the ICU.
Questions from organizations
What organizations usually ask.
What data does Corpus need to run the model?
Clinical data that already exists in the system: lab results, vital signs, diagnoses and medications. The published model was derived from routine clinical data in electronic health records. The data comes in through HL7, FHIR, CSV or files exported from the health record; Corpus structures and deduplicates it before calculating risk.
How long does an implementation take?
It depends on how available and complete the clinical data already in the organization's system is. The model is not retrained for each client: it is recalibrated to the local incidence, as the article recommends, and validated before the cohorts are turned on.
Does the model work outside Latin America?
There is an external validation. The same model, without retraining, was validated in NIH All of Us with 626,396 US adults and obtained O/E ≈ 0.98. That validation is separate from the IJCRP 2026 article, which describes the model's derivation in 382,589 patients in Colombia.
How is patient data protected?
We process personal data under Colombia's Law 1581 of 2012, and we comply with HIPAA (United States) and GDPR (European Union) to protect health data. The specific security measures and the terms of processing are set with each client in the implementation agreement.
What does the clinical team receive?
Actionable cohorts, not a score. Each patient lands in one of four cohorts, with its clinical action, its time window and the modifiable factors that explain their risk. That way the team knows whom to escalate within 24 hours, whom to contact within 7 days, whom to give a care plan and whom to follow passively.
How is the return measured?
By events prevented and cost avoided against the client's baseline. In a real production deployment with a health-system partner in Latin America: 550+ acute cardiovascular events prevented, $2.3M USD saved and ↓20% monthly heart attack incidence. Name withheld for confidentiality.
Next step
Want to see the platform with your organization's data?
30 minutes with our team. We look at your case.