AI-Backed Medical Technology Can Augment, Refine Diagnosis

Q: How does Medsi aim to transform the way individuals and institutions approach healthcare monitoring and diagnostics?
A: We are a digital medical platform powered by proprietary medical reasoning algorithms. We remotely monitor a wide range of vital signs and biomarkers. We use this information to generate insights that allow us to identify opportunities for preventive care and early intervention. We ensure continuous monitoring and follow-ups, guiding patients through personalized lifestyle changes and targeted interventions that help them manage certain chronic conditions.

Q: What makes your healthcare solution particularly relevant and viable in the Mexican market?
A: We leverage the device that everyone carries in their pocket: a smartphone. We can use virtually any smartphone. This gives us the ability to begin driving meaningful change.
Our business model focuses on B2B and public sector partnerships. Budget cuts and other structural issues have placed additional pressure on an already overburdened system. This problem, combined with demographic shifts and a growing population demanding more healthcare services, means that the adoption of technology is no longer optional.

Q: What technical and regulatory challenges did Medsi face when obtaining COFEPRIS’s approval, and how did the team overcome them?
A: We achieved an unprecedented feat: introducing the first solution to authorization as a Class II medical device software (SaMD), not only in Mexico but throughout the Western Hemisphere, with a sanitary registration from COFEPRIS. Challenges stem primarily from innovation. The processes and criteria established by COFEPRIS to protect patients were designed for physical devices. A large part of the challenge was figuring out how to adapt requirements originally designed for physical products to software. We had to work together closely with authorities to help them understand why our solution was fundamentally different. We also had to demonstrate that our technology is just as effective and offers accuracy comparable to medical-grade equipment across all population groups.

Our device was also validated to work across a broad range of skin tones and body mass indexes. It maintains the same level of precision whether the subject has a beard, is clean-shaven, wears a mask or makeup. Additionally, we’ve validated our efficacy with large-scale clinical trials on local population. This involved performing testing on thousands of patients across Mexico following strict protocols.

Q: How does your technology work, and what makes it a powerful diagnostics tool?
A: The technology itself is similar to that used in devices like the Apple Watch. The key difference is in the source of light. Devices like an Apple Watch, a Fitbit, or an Oura ring all emit a beam of light onto the skin. We extract three beams of visible light at different frequency ranges that penetrate the hypodermis to varying depths. We then map a handful of blood vessels near the surface of the face’s skin. We monitor these vessels by measuring the reflection and absorption of light signals, effectively recreating what happens in the blood vessels beneath the skin.

The device captures 30 frames per second across three light beams, collecting 90 data points per second over 70 seconds. This results in 6,300 precise data points. The real value lies in how the raw data is transformed into actionable health information.

Q: What sets your platform apart from other remote health monitoring solutions, and how do you turn data into meaningful medical insights for both patients and physicians?
A: Remote monitoring is just the starting point. The true value comes from what we can accomplish with all that information. Our company has developed specialized medical reasoning algorithms powered by AI and designed specifically for healthcare applications. We have created a proprietary RAG (Retrieval Augmented Generation) layer to train these AI models to perform specific medical functions. This allows us to compile a comprehensive health profile for each patient in about five minutes, which includes clinical history, family background, lifestyle, mental health, and results from any lab, imaging or clinical tests. When this data is combined with a long history of vital signs readings, the system can generate illuminating and highly accurate medical inferences.

Among other things, our algorithms generate a variety of personalized health reports, both for end users and for health workers. For patients, the platform translates findings into clear, everyday language, providing detailed interpretation and explanations, highlighting lifestyle interventions, possible deficiencies, hereditary risks, and suggesting relevant lab tests or referral to specialists. For doctors, the output is clinical and in-depth, identifying abnormal variables, suggesting potential diagnoses, and analyzing medication interactions. Once we’ve generated these personalized health plans, we guide patients through a variety of lifestyle interventions with continuous monitoring, improving their adherence to treatment.
Overall, this approach allows for more personalized and proactive healthcare, improving patient understanding and aiding physicians in making better-informed decisions.

Q: How does the platform facilitate data sharing between patients and healthcare providers?
A: We have a feature we call the Health Wallet that functions as a personal medical folder, securely storing a patient’s entire medical history including records, medications, vital signs readings, lab or imaging test results, and more. This information can be shared with any chosen doctor, specialist, or primary care physician.

Q: How is Mexico’s shifting demographic profile impacting the country’s healthcare needs, and how does your solution address these challenges?
A: Mexico’s population is aging. Over the past 40 years, the country shifted from being primarily a pediatric market to a mature one. As a result, the prevalence of chronic degenerative diseases has increased significantly, but Mexico’s healthcare system was not designed to handle this shift. There are also socioeconomic and ethnic factors to consider. For example, certain ethnic groups have a higher predisposition to metabolic syndrome and diabetes. Moreover, about 55% of health spending in Mexico is paid out of pocket. Mexico has the second-highest out-of-pocket health expenses in the world, second only to India. This means the financial burden of this epidemiological challenge falls heavily on families.

The public health sector also faces significant limitations and constraints. Over 50% of people with type 2 diabetes are unaware of their condition. Similarly, over 65% of those with hypertension do not know they have it. Our first goal is to address underdiagnosis through large-scale population screening campaigns, allowing early detection before these conditions lead to catastrophic health events. Economically, preventive care costs about 5% of what reactive care costs once a crisis occurs. Once we’ve identified this population at risk of chronic diseases, we can help them introduce lifestyle changes to manage (or, if caught early enough, even to reverse) these conditions.

Q: How does your solution help reduce strain on the healthcare system and support medical professionals, particularly in underserved or remote areas?
A: We will leverage smartphones to overcome geographical barriers that limit access to healthcare services. Our solution addresses several of the key bottlenecks across a patient’s entire healthcare journey. By starting with a remote vital sign scan, we can perform triage and successfully address queues at public sector facilities, increasing the healthcare system’s capacity and efficiency. Patients will also arrive at the doctor with a complete set of medical records and measurements of critical health parameters. The public healthcare system faces a severe shortage of personnel. Using diagnostic tools that analyze all this information to rule out certain conditions helps immensely. We do not aim to replace doctors but to support them.
Finally, once someone has received treatment, our platform can perform long term continuous monitoring and follow-up remotely, further reducing the strain of long term care on the healthcare system.

Q: How do you ensure the security and confidentiality of users’ personal health information?
A: Our solutions are “edge-based”, meaning that all processing is done within the device itself. Data is only sent to our servers when additional materials – like lab results, imaging studies or pictures of external symptoms – are uploaded into the platform for interpretation. Patients’ health information remains safely locked in their own device.

We have been conscientious and deliberate to ensure compliance of all privacy comply with all regulations. Great value is derived from processing large amounts of anonymized information, which do not involve Personally Identifiable Information (PII), so it is unnecessary for us to use any individual patient’s data to extract demographic, epidemiological, and pathological trends from large populations. This information guides public policy and helps allocate budgets more effectively.
We collaborate with public entities and give them access to a dashboard where they can monitor everything that is happening across their territory, but always in an anonymized manner.

Q: What strategic motivations led to your alliance with Laboratorios Sanfer to raise US$3 million in a Seed 1 round?
A: During our first 3 years in operation, Medsi.ai was bootstrapped by founders. Thus, our goal in bringing new investors onboard was not merely to raise funding but to align ourselves with the smartest money in Mexico’s health industry. Laboratorios Sanfer has been a leader in the Mexican pharmaceutical space for over 80 years. No one is more in tune with the health industry and its challenges than the company. Beyond the funding, the real value lies in its support and guidance in key strategic decisions. For example, in matters such as protection of our intellectual property and navigating the regulatory pathways, having access to Laboratorios Sanfer’s expertise is invaluable. Their experience and connectivity greatly benefit us in terms of access to commercial partners and more.

Q: What new features or tools are you planning to add to the app in the near future, and how will they enhance diagnostic capabilities and preventive care?
A: We are launching several new features in our platform over the next couple months. One key tool is a symptom check feature, where users can upload photos of any visible symptoms along with a brief description of symptoms and answers to relevant questions. Our medical reasoning algorithms analyze this data – along with the rest of the patient’s health profile – to estimate the probability of various diagnoses with remarkable accuracy, overcoming a major limitation of remote diagnosis: the inability to perform physical exams.

We are also expanding the range of our non-invasive, “bloodless” blood tests that presently measure hemoglobin and glycated hemoglobin (HbA1c), and plan to add six more metrics within the next year, including cholesterol, fasting glucose, albumin, and globulin. Our goal is to provide additional blood chemistry analysis through the ou remote,contactless measurements. We are working closely with the government on large-scale screening and detection campaigns to obtain a clear picture of the population’s health. From there, targeted interventions can be implemented for groups facing specific conditions. Early results from thousands of users show that after six months, 40% of users experience improvements in their state of health.

Medsi AI is a platform that converts any smartphone, laptop, or tablet into a diagnostic and monitoring tool with medical-grade accuracy.