Artificial intelligence shows a pattern in healthcare: overpromise a replacement, underdeliver on integration, and leave the clinical relationship worse off than it was before. Myra Ahmad, CEO of Mochi Health, argues this is avoidable. Her view, discussed on the Compound Wisdom podcast in April 2026, is that AI is powerful for backend operations but not yet ready to replace physician judgment.
Most providers struggle with fragmented systems where labs, medications, and treatment plans sit in disconnected silos. This is the kind of problem AI handles well. Pattern-matching across large volumes of data, flagging a lab trend, or surfacing drug interactions benefits from a system that never gets tired.
Routing patients to the right provider, catching scheduling conflicts, or streamlining clinical oversight work behind prescriptions requires speed and consistency at scale. That is where AI’s advantage over human systems is most immediate and least controversial. The technology handles the unglamorous work that keeps a marketplace moving.
The Limits of Algorithmic Care
Ahmad’s argument against AI in clinical decision-making is grounded in the specific challenges of personalized care. In a MedCity News op-ed, she noted that individualized GLP-1 dosing “demands greater medical supervision, not less.” An algorithm trained on an average patient reproduces the exact failure of rigid clinical protocols: models built around a trial population that does not represent the person being treated.
Handing that decision to AI without a physician in the loop automates bias at scale. There is also the issue of trust. A patient disclosing a side effect they are embarrassed about or explaining why they stopped taking a medication provides context a model trained on structured data cannot read. The most consequential clinical decisions happen where data is incomplete and judgment is required, not where a pattern is already obvious enough for software to catch.
Providers on Mochi Health practice without interference from a system dictating outcomes to them. The platform acts as infrastructure rather than a decision-maker. This division of labor keeps the physician firmly in charge of individual choices while letting AI handle the repetitive logistical tasks that often cause patients to fall out of care.
Ultimately, the industry is still oscillating between AI hype and AI panic. Ahmad’s position offers a specific line: let the technology do what it is actually good at and keep the physician in charge of the rest. That is not a hedge against AI, but an argument for building with it correctly from the start.
Adopting a hybrid model that prioritizes human expertise over automated precision creates a more sustainable path forward. By treating the relationship between doctor and patient as the core unit of care, the industry can ensure that AI serves the patient rather than the other way around.
This approach aligns with broader discussions on health optimization, where lifestyle choices and medical interventions intersect to improve long-term well-being. The goal is a cohesive strategy rather than a series of disconnected fixes. [1] Lifestyle choices and medical interventions intersect to improve long-term well-being.