About 1 in 8 US adults are now on a GLP-1, and a patient can get a prescription tonight without ever speaking to a clinician. Fill out a form, pay the subscription, and a vial ships to the door. This creates an almost irresistible opportunity for your patients.
The work is now deciding whether the drug is helping or harming your patient. Is the patient preserving lean mass? Are their biomarkers improving? Are they taking GLP1s for reasons other than weight loss and metabolic health? Are they building habits so they can sustain the results off the drug?
Tracking all of these variables used to be a bottleneck because medication management is inherently high-touch, complex clinical work, and AI is exactly how you manage it at scale.
The market is creating the gap between script and care
Cost, not side effects, is one of the most common reasons patients abandon GLP-1 therapy. If you look at how cheap DTC GLP-1s are, you can now see why patients pick them over the doctor’s office. Hims & Hers lists oral weight loss kits around $79 a month and compounded GLP-1 injections around $199 a month, well under the cash price of branded semaglutide or tirzepatide. The results are real, too. In one analysis of more than 50,000 telehealth patients, average weight loss reached 8.9% at three months and 19% at twelve months, in line with the major trials. So this is not a story about fake drugs or fake outcomes.
The market is sorting into two models. On one side, low cost platforms sell access and autonomy. On the other, a wave of better funded players is racing to wrap medication in real support. Nourish, a dietitian led metabolic care company, just raised $100 million to pair GLP-1 prescribing with registered dietitians, lab testing, and AI engagement tools. Others like FuturHealth bundle dietitian designed nutrition plans with the meds. Legacy consumer brands are pivoting, too. Noom has expanded into a preventive health platform, offering a GLP-1 "microdosing" strategy that pairs low-dose prescriptions with at-home biomarker testing, muscle defense, and behavioral coaching. Investors have figured out that the prescription is the entry point, and the care is the competitive advantage.
Telehealth platforms are built to start the medication, not stop it. Without clinical oversight to build sustainable habits, the patient is simply renting a result until the subscription ends.
The clinical risk hiding in the convenience
Here is the true cost of a cheap prescription. A GLP-1 suppresses appetite so effectively that, without clinical oversight, patients often deprive themselves of the nutrients required to build muscle and sustain their results. A 2026 scoping review in Obesity Reviews of twelve trials found that while patients cut calories by up to 39%, roughly 40% of the weight they lost was lean tissue. Shockingly, only three of those twelve studies even involved a registered dietitian. Stripping away that much muscle drastically raises the risk of osteopenia, falls, and a slowed metabolic rate that practically guarantees future weight regain.
And that regain is the other half of the crisis. In a cohort of more than 125,000 patients, nearly two thirds of those without diabetes stopped taking the drug within a year. Once the medication stops, the clinical reversal is staggering:
Patients regain about 60% of the weight within one year, with continued regain over time.
Blood pressure, lipids, and glycemic benefits return to baseline within 12 to 18 months.
Quitting within the first year raises one-year coronary artery disease incidence to 17.1% (from 13.2% on treatment) and heart failure to 10.2% (from 7.3%).
Clinicians are feeling this widening care gap in the off-ramp. Registered dietitian Sarah Skobeloff shared, "You can't have interdisciplinary team care (PCP, cardiologist, endocrinologist) when using an online company." Endocrinologist Marc-Andre Cornier, former president of The Obesity Society, has warned that many telehealth companies don't properly evaluate patients before writing the prescription. The solution? A joint advisory from four major obesity and nutrition societies lays out the basics:
Protein at 1.2 to 1.5 grams per kilogram per day
Resistance training
Ongoing monitoring for nutrient deficiencies
The guidelines are straightforward, but the math of delivering them isn't. An app can scale a cheap prescription in seconds, but high-touch clinical care takes time. To catch up to the telehealth machine, you just need a way to see what your entire patient panel is doing between visits. That is where a clinical AI playbook comes in.
Beyond the Scale: The AI Playbook for Total Metabolic Tracking
The newest clinical AI models don't just log data; they synthesize it. By reading body composition reports, CGM data, and food and exercise journals, they connect the dots and amplify clinical signals hiding in the noise of all the chart documents in your task list to review.
Picture a patient three months into semaglutide. Her DEXA scan shows 14 pounds down, but 6 of those pounds are lean mass. Her continuous glucose monitor reads flat and low because she is barely eating. Her food photos confirm it: coffee, a few bites at dinner, almost no protein. Each data point looks fine alone. Together they show a patient undereating her way into muscle loss. A clinical AI tool lines up all three in seconds and surfaces the pattern, so you start a protein target and a resistance plan before the next scan. Here is how to build each stream:
The body composition stream. The scale and BMI are not simply good enough anymore. You need to see the body composition. DEXA is the gold standard for precision, and mobile clinics like BodySpec make it incredibly accessible for your patients to get baseline and quarterly scans. For higher frequency tracking in the clinic, medical grade bioimpedance (BIA) devices like SECA provide fast monitoring, though you must account for hydration drift. Use AI to track the lean mass and visceral adipose tissue trends across any of these scans, instantly flagging the patient who is losing muscle too fast.
The glucose stream. An A1C gives you a snapshot, but OTC monitors give you a continuous window. With the FDA clearing Dexcom’s Stelo and Abbott’s Lingo, getting patients on a sensor no longer requires a prescription. But raw CGM data is just noise without context. Feed that stream into an AI tool alongside a food log. The model cross-references the two to show you exactly which meals drive a spike. More importantly for your GLP-1 patients, it also can catch the opposite pattern: the flat, low curve that confirms they are not eating enough.
The food log stream. Patients log meals as photos and plain text, and that used to pile up unread. Clinical AI tools now estimate nutrients from a picture of a plate. Research frameworks like DietAI24 identify foods and compute dozens of nutrients from real world images, and commercial engines like Passio offer the same through an API. Keep the caveat in view: head to head tests show general models like GPT-4o, Claude, and Gemini still vary widely on portions and calories, so use the estimate to catch the patient undereating protein, then confirm before you counsel on numbers.
Feed these streams into your clinical AI tool or AI-native EHR like Ultralight. Purpose-built for medicine, it accurately extracts data from messy PDFs, lab reports, and scans. It does more than just display raw numbers. It tracks longitudinal changes and amplifies the clinical signals hiding in the noise. By surfacing exactly what you need for your next clinical decision, you align the therapy, the provider, and the outcome. That is a standard of care a website cannot provide.
Before you trust the output
One discipline carries over from our genomics issue last week. An AI tool is only as good as the model behind it and the data handling around it. AI can still hallucinate. One reading a food log can invent a nutrient total, and one reading a scan can state a confident number that is wrong. Treat every output as a draft you verify, not a result you sign.
For anything that touches identifiable patient data, confirm three things first:
The vendor has a signed BAA
You know their data retention and training policy
The tool runs on a compliant deployment rather than a shared consumer endpoint
A general purpose chatbot in its consumer form is not the place for identifiable labs or images. A purpose built clinical tool with HIPAA compliant architecture is a different category. Know which one you are using.
What to try this week
You do not need to rebuild your practice to close the biggest gap on a GLP-1. Pick one of these AI assisted plays and run it with five patients.
Put a nutrition coach in their pocket. Ask each GLP-1 patient to photograph their meals and run them through a custom GPT that reads the plate and gives instant nutritional feedback. Here is one you can hand to patients today: Lexi’s Nutrition Companion on ChatGPT. It scores their protein, flags when they are undereating, and keeps a running log they can show you at the next visit. The patient gets coaching between appointments, and you get a week of real intake instead of a vague recap.
Walk into every visit already briefed. Point an AI summary at each patient's incoming data, the meal photos, check in answers, weight, and labs, and have it draft a one paragraph pre visit brief that flags who is losing muscle, undereating protein, or sliding toward a deficiency. You spend the first two minutes of the appointment acting on a synthesis instead of hunting for it, and nothing important slips past.
Then widen it. The patient who bought their script online is unmanaged, not unreachable. A short message offering a body composition baseline and a nutrition check is the easiest new patient conversation you will have this month.
Be a modern clinician with the help of Ultralight, the AI-native EHR built specifically for functional, integrative, and longevity medicine.
In the news
Dietitian led GLP-1 care just got a $100M vote. Nourish raised $100 million to scale its AI assisted, dietitian anchored metabolic platform. The investment thesis is the same one this issue makes: the script is cheap, the wraparound care is the value.
Harvard just dropped its first-ever longevity report. Harvard Medical School and Harvard Health Publishing released a Special Health Report on Longevity covering where GLP-1s, rapamycin, metformin, senolytics, and emerging peptides stand in the current evidence. The report's framing of GLP-1s as a longevity therapeutic, not just a weight-loss drug, is the cultural moment to watch. Your patients will start asking how their script fits a broader longevity protocol.
Oura just bundled an AI doctor and a real physician into a ring. Oura launched the Ring 5 with a Counsel Health integration, letting members chat with an AI doctor inside the app and escalate to a licensed physician if needed. The feature is live in 43 states and includes GLP-1 management. Your patients will start arriving with AI consults already in hand, the same way they arrive with chatbot-generated lab interpretations today.
Upcoming Conferences & Events
Jun 9–11, Longevity Docs Cannes 2026 · Cannes, France · Invite-leaning room for clinicians at the frontier of longevity medicine. Worth it if you are designing the next iteration of your own practice.
Sep 24–26, Vibrant Longevity Summit · Austin, TX · A clinical room of practitioners running lab-driven, longitudinal care. For anyone building a practice around diagnostics and biomarkers who wants peers who work the same way.
Oct 8–10, A4M Women's Health Summit · San Antonio, TX · The best clinical education on hormone, metabolic, and midlife women's health you will see this year. The room to be in if you are growing the perimenopause and menopause side of your practice.
Oct 21–24, NAMS Annual Meeting · San Diego, CA · The single most practice-changing meeting of the year for midlife women's health. Your protocols will look different after this one.
Nov 5–8, Eudēmonia Summit · West Palm Beach, FL · One of the most talked-about longevity gatherings in the U.S. Experientials, hands-on demos, and the best place to try the emerging frameworks your patients will ask you about next year. Ovation and Ultralight team will be there!
Nov 5-7 — Private Physicians Alliance Annual Meeting · St. Petersburg, FL The gathering for independent, cash-pay, and concierge physicians navigating practice independence. Practical and peer-driven. Ultralight will be there!
Nov 8-11 — American College of Lifestyle Medicine Conference · Orlando, FL Lifestyle medicine's main annual event — evidence-based approaches to behavior change, chronic disease, and healthspan. Growing overlap with the longevity medicine community.
Dec 11–13, A4M Longevity Fest · Las Vegas, NV · The biggest longevity event in the U.S. The room spans clinicians, industry, founders, and the people building next year's platforms, and the connections from this one tend to compound through the rest of your year. Ultralight will be there!
Know of an event we should add? Reply and tell us.
Until next week
The prescription is solved. The year after is not. That year is where your patients keep their muscle, their bone, and their results. AI doesn't replace that work. It makes it fast enough to deliver for every patient, not just the one in front of you today.
Pick one GLP-1 patient this week. Get a body comp baseline, run their food log, and find one thing worth acting on. Reply and tell us what you find, or what is getting in the way. The best ideas in this newsletter come from clinicians doing the work. Browse past issues in the archive if you missed one.
Until next week, keep building the practice you imagined when you started. We are building it with you.
— Sunita and Dr. G