For the patient sitting in an endocrinology clinic with a complicated thyroid case, perimenopausal symptoms, or a request for a peptide regimen they read about online, the clinical question is rarely simple and the literature behind the answer is rarely settled. Most prescribers do not have time during a fifteen-minute visit to read three randomized controlled trials, weigh a regulatory update from the FDA, and grade the strength of the evidence. Hormonaly.ai, a healthtech startup led by Canadian endocrinologist and geneticist Fady Hannah-Shmouni, is betting that 24 specialized AI agents can do that synthesis work in under 100 milliseconds and hand the clinician a cited, GRADE-rated answer before the patient finishes describing their symptoms [Hormonaly.ai].
The patient population in question is broad and growing: adults navigating hormone-related conditions including hypothyroidism, hypogonadism, polycystic ovary syndrome, adrenal disorders, and the increasingly common request for peptide therapeutics such as GLP-1 receptor agonists, growth hormone secretagogues, and other compounded agents. Many of these patients today are seen across a fragmented care landscape: primary care offices that may not specialize in endocrine nuance, telehealth clinics that have grown up around testosterone and weight-loss prescribing, and a smaller number of board-certified endocrinologists with multi-month wait times. Standard of care typically involves serum hormone panels, a clinical history, and treatment guided by society guidelines from groups like the Endocrine Society or AACE, with adjustments based on the prescriber's experience. Peer-reviewed evidence on newer peptides is uneven, and compounded products sit in a regulatory gray zone that the FDA has been actively reshaping over the past two years.
Into that gap, Hormonaly is positioning what it calls an "evidence OS for modern medicine," with an AI copilot focused specifically on peptide and hormone medicine [Hormonaly.ai]. The company says its platform synthesizes from more than 10,000 PubMed studies and delivers 631 GRADE-rated protocols through 24 agents purpose-built for clinical workflows [Hormonaly.ai; LinkedIn]. The pitch to a prescriber is straightforward: instead of opening a browser tab to PubMed mid-visit, ask the agent and receive an answer with real-time citations attached [Hormonaly.ai].
The bet
The wedge Hormonaly is choosing matters. Rather than competing with general-purpose clinical AI tools that try to cover every specialty, the company is going narrow: hormone and peptide medicine, where prescribing volume is rising sharply, where guidelines lag the literature, and where clinicians are often working without a specialist nearby. GRADE, the framework Hormonaly says it uses to rate its protocols, is the same methodology that underpins guidelines from the WHO and most major medical societies, which gives the output a vocabulary that practicing clinicians already trust [LinkedIn].
| Metric | Value |
|---|---|
| PubMed studies referenced | 10000 studies |
| GRADE-rated protocols | 631 protocols |
| Specialized AI agents | 24 agents |
Why it could matter
The tailwinds are real. Demand for hormone-related care has surged alongside the GLP-1 era, and the FDA has spent the past 18 months tightening its posture on compounded semaglutide and tirzepatide as the branded shortages resolved. Clinicians prescribing in this space are operating under genuine medico-legal pressure to document evidence-based reasoning. A tool that produces a cited, graded protocol at the point of care is, at minimum, a defensible workflow artifact. If the underlying synthesis holds up to scrutiny from practicing endocrinologists, the addressable user base extends well beyond specialty clinics into primary care, men's health platforms, menopause-focused telehealth providers, and longevity practices.
The broader thesis Hannah-Shmouni has articulated in podcast appearances is a shift from reactive "sickcare" to proactive, personalized health management, with AI bridging the gap between practitioners and the evidence base [Leaders On Purpose Podcast; PHAMCAST]. That framing is consistent with where reimbursement and patient demand are pulling hormone medicine, particularly around healthy aging and metabolic health.
The team
Hannah-Shmouni's clinical credentials are unusually deep for a healthtech founder. He is an MD, FRCPC endocrinologist and geneticist, a clinical professor of medicine at the University of British Columbia, and previously served as associate program director of the NIH's Inter-Institutes Endocrinology and Metabolism Fellowship Program [LinkedIn; Erdheim-Chester Disease bio]. At the NIH, he was a principal investigator on two NICHD protocols covering primary aldosteronism in Black patients and creatine transporter deficiency in males, and he led work in the Stratakis Laboratory's section on endocrinology and genetics [Erdheim-Chester Disease bio]. He has authored more than a hundred scientific publications [Mizter Rad Show] and was named to the medical advisory board of Science & Humans, a Canadian hormone-focused telehealth provider [Yahoo Finance]. He is also Managing Partner at FHS Capital [Erdheim-Chester Disease bio]. For a product whose entire credibility rests on the quality of its evidence synthesis in endocrinology, having a working endocrinologist with an NIH research record at the helm is a meaningful asset.
The honest counterfactual
What skeptics will rightly want is peer-reviewed validation. A platform that promises GRADE-rated protocols in under 100 milliseconds is making a strong claim about the fidelity of its retrieval and reasoning, and to date the public evidence for the system's accuracy comes from the company's own materials [Hormonaly.ai]. Hallucinated citations have been a documented failure mode of general-purpose LLMs in clinical settings, and a tool that cites PubMed convincingly but incorrectly could create real downstream risk for prescribers and patients. The bull answer is that Hormonaly's narrow domain (hormones and peptides) and its grounding in a curated literature corpus are precisely the conditions under which retrieval-augmented systems perform best, and that founder-led clinical oversight from someone with Hannah-Shmouni's publication record raises the floor on quality control. An independent prospective evaluation, ideally published, would convert that argument from plausible to demonstrated.
What to watch
Over the next twelve months, the milestones that would move the story forward are concrete: a published validation study comparing Hormonaly's protocol outputs against expert endocrinologist consensus; named pilot deployments inside hormone-focused telehealth providers or academic endocrine clinics; a disclosed funding round that signals institutional conviction; and clarity on whether the product is sold to individual clinicians, to clinics as a SaaS seat, or embedded into EHR workflows. The category Hormonaly is reaching for, point-of-care evidence synthesis for a high-volume specialty, is one of the more defensible applications of clinical AI on offer. Whether this team is the one to define it depends on what the next study, and the next set of users, actually find.
Pulse Raman covers biotech, digital health, and clinical AI for Startuply.