Ro's AI Strategy: How the Telehealth Pharmacy Is Rebuilding Patient Intake in 2026

TL;DR#

Ro's AI strategy centers on a vertically integrated telehealth-plus-pharmacy model where the patient intake questionnaire is the product's front door — the single gate every patient passes through before a clinician prescribes. Ro hit roughly $598M in annualized revenue in 2024, up 66% year-over-year, with about $370M of that coming from its GLP-1 weight-loss business, according to Sacra's estimates. The company has treated millions of patients since 2017, reaching one in every U.S. county and 98% of primary-care deserts, and was last valued at $7 billion. Ro's most concrete AI move is operational: a large language model that triages GLP-1 adverse-event messages with 94% accuracy and cut mean care-team response time from 115.1 minutes to 33 minutes, presented at ObesityWeek 2025. But the static intake form that opens every Ro visit still can't probe the hesitation, history, and "why now" that determine whether a patient stays on therapy — and up to 57% of GLP-1 patients discontinue within six months without behavioral support. Conversational AI intake closes that gap by interviewing patients in their own words at the exact moment they decide to start. For Perspective AI's view on why this matters, see why AI-first research can't start with a web form.

What is Ro's AI strategy?#

Ro's AI strategy is a vertically integrated approach that uses large language models to automate clinical triage and operational workflows across its telehealth, pharmacy, and diagnostics stack, while keeping a static digital intake questionnaire as the entry point to every patient relationship. Ro (formerly Roman Health Ventures, led by co-founder and CEO Zachariah Reitano) is a direct-to-consumer healthcare company built around men's health, women's health, and — most recently — GLP-1 weight management, where the online visit form gates eligibility before a licensed clinician reviews and prescribes.

The strategically important part is structural. Because Ro owns the telehealth platform, the pharmacy network, and in-home diagnostics — unified under what it calls ro.OS, launched in 2024 — the intake questionnaire isn't just a lead-capture form. It's the moment Ro decides who qualifies for care, what they're prescribed, and how the relationship begins. That makes intake the highest-leverage surface in the entire business, and the one most exposed to the limits of static forms.

Ro at scale: a telehealth pharmacy built on intake#

Ro operates one of the largest direct-to-consumer telehealth pharmacies in the United States, with intake volume that rivals a national clinic network. Five numbers frame the scale:

The reacceleration was driven almost entirely by Ro Body, the GLP-1-centered obesity program that became the company's primary growth engine by late 2023. That shift matters for intake: weight-loss patients arrive with more complex histories, more hesitation, and far higher long-term adherence stakes than a patient renewing a refill. Every one of them enters through the same questionnaire.

Ro has also positioned itself as a distribution partner to pharma rather than an adversary. It launched the Wegovy pill through an integration with Novo Nordisk, and when Novo Nordisk ended its branded Wegovy relationship with Hims & Hers in June 2025, it kept its deals with Ro and LifeMD intact. For how a competing telehealth brand approaches the same intake problem, see our Hims & Hers AI patient intake case study.

Where Ro uses AI today#

Ro's most concrete AI deployment is an LLM-based clinical triage tool, not a marketing chatbot. At ObesityWeek 2025, Ro's research team presented an evaluation of a large language model designed to automatically detect and triage patient messages containing GLP-1-related adverse-event concerns. The results were specific and operationally meaningful:

• The model achieved 94% accuracy in detecting adverse-event-related message content.
• After implementation, mean care-team response time fell from 115.1 minutes to 33 minutes — a roughly 71% reduction.

This is AI doing exactly what it should in healthcare: routing the right message to the right clinician faster, without replacing clinical judgment. Ro presented eight research abstracts at ObesityWeek 2025 in total, several explicitly advancing how AI and LLMs can support the patient experience.

Ro's clinical outcomes give the model something worth protecting. The company published the first study evaluating long-term weight-loss and safety outcomes for semaglutide delivered via telehealth, reporting that patients lost an average of 16.6% of body weight over 68 weeks — outcomes consistent with the original clinical trials. The strategic question is whether the front door — intake — is feeding that engine the context it needs. This is the same tension we mapped in the Teladoc telehealth network case study: strong AI on the clinical side, static forms on the intake side.

Why form-based intake bottlenecks patient understanding#

Form-based intake bottlenecks patient understanding because a fixed questionnaire collects the fields a clinician needs to prescribe but not the reasoning that predicts whether a patient will stay on therapy. A Ro online visit asks about health, lifestyle, medical history, and symptoms — structured, branching, and clinically sound. But branching logic is still a decision tree the patient walks down, not a conversation that follows them.

The cost of that gap shows up directly in GLP-1 adherence. In a peer-reviewed analysis of direct-to-consumer telemedicine, up to 57% of GLP-1 patients discontinued therapy within the first six months without ancillary behavioral and adherence support. The recorded reasons were revealing:

• Program cost — 38.7%
• Dissatisfaction with results — 16.8%
• Intolerable side effects — 15.4%
• Reached weight-loss goal — 7.3%
• Temporary pause — 5.6%

Almost none of those discontinuation drivers are knowable from a yes/no intake field. "Will this fit my budget once insurance changes?" "I tried a GLP-1 before and quit — will this time be different?" "I'm nervous about nausea." A static form has no way to ask "why now?" or to follow up when an answer is vague. It captures eligibility; it misses the hesitation, prior-attempt history, and motivation that determine adherence and trust. We unpack this failure mode in detail in AI vs. surveys: why conversations win for real customer research.

The broader market backdrop reinforces the stakes. McKinsey estimated that up to $250 billion of U.S. healthcare spend could shift to virtual or virtually enabled care, with virtual care now running at 14–17% of visits versus roughly 1% before the pandemic. As more of the patient relationship moves online, the intake form becomes the dominant first impression — and the place where understanding is either captured or lost.

What conversational AI intake unlocks for telehealth#

Conversational AI intake unlocks the reasoning behind a patient's answers by interviewing them in natural language — asking follow-ups, probing uncertainty, and capturing the "why now" — instead of forcing them into dropdowns. For a GLP-1 program like Ro's, that turns intake from a compliance checkpoint into the richest adherence signal the company collects.

Consider the difference in three real intake moments:

1. Prior-attempt history. A form asks "Have you taken a GLP-1 before? Yes/No." A conversational agent hears "yes," then asks why the patient stopped, what side effects they hit, and what would make this attempt stick — surfacing the 15.4% side-effect risk before it becomes a discontinuation.
2. Cost sensitivity. A form has no field for "I can afford this for three months but I'm worried about month four." A conversation does, and routes that patient to support before cost (the single largest discontinuation reason at 38.7%) ends the relationship.
3. Motivation and "why now." A form can't tell a wedding from a recent diagnosis. A conversational interview captures the motivation that clinicians and care teams can reinforce.

This is precisely the gap Perspective AI is built to close. Our AI interviewer agent conducts hundreds of natural-language conversations simultaneously, and our concierge agent replaces the static intake form with a conversation that still collects every required field — while capturing the context a form discards. For telehealth specifically, Perspective's intelligent intake is designed to gate eligibility and capture reasoning in the same flow. The same pattern is reshaping onboarding across regulated industries, as we documented in the Chime AI onboarding case study and the Lemonade conversational-AI insurance case study.

Ro customer research: from intake form to continuous voice of patient#

Ro customer research today runs largely off structured intake data and post-visit questionnaires, which means the company is rich in clinical fields but comparatively thin on the unprompted "why" behind patient behavior. A 50-day follow-up questionnaire is a useful checkpoint, but it's still a survey — and surveys inherit every limitation of the intake form that preceded them.

The opportunity is to treat intake as the start of a continuous voice-of-patient program rather than a one-time gate. When the front door is a conversation, every patient interaction generates qualitative signal that clinical teams, product teams, and marketing can act on. A churn spike tied to side effects, a pricing objection that recurs across thousands of conversations, a competitor mention — these surface in days, not in the next quarterly survey cycle. Our guide to building a voice-of-customer program from scratch lays out the operating model, and the 2026 customer-research tool stack maps where conversational intake fits. Telehealth teams running this play are functionally CX teams and product teams at once — they own both the clinical relationship and the funnel.

The healthcare context: why this matters in 2026#

In 2026, the telehealth companies that win GLP-1 and chronic-care patients will be the ones that understand patients at intake, not just qualify them. Ro's AI strategy already proves the company can deploy LLMs responsibly on the clinical side — 94% triage accuracy and a 71% cut in response time are serious results. The unfinished work is on the front door, where a static questionnaire still stands between a hesitant patient and an adherence program that demonstrably works.

This isn't a Ro-specific problem; it's a category problem. As we noted in our analysis of how Maven Clinic approaches women's-health telehealth onboarding and how One Medical is modernizing patient onboarding, the entire industry has invested in AI everywhere except the moment a patient first describes their own situation. Whoever fixes intake first owns the adherence advantage.

Frequently Asked Questions#

What is Ro's AI strategy in healthcare?#

Ro's AI strategy uses large language models to automate clinical triage and operational workflows across its vertically integrated telehealth, pharmacy, and diagnostics platform. The clearest example is an LLM that detects GLP-1 adverse-event messages with 94% accuracy and cut mean care-team response time from 115.1 minutes to 33 minutes, presented at ObesityWeek 2025. Ro's patient-facing intake, however, still relies on a static online questionnaire rather than conversational AI.

How does Ro telehealth patient intake work?#

Ro telehealth patient intake works through a dynamic online visit in which patients answer branching questions about their health, lifestyle, medical history, and symptoms, after which a licensed clinician reviews the responses and prescribes medication if appropriate. The questionnaire is the single gate every patient passes through. It is clinically sound for determining eligibility but, like any form, it captures fields rather than the reasoning, hesitation, and motivation that predict long-term adherence.

How big is Ro and what is it valued at?#

Ro reached roughly $598 million in annualized revenue in 2024, up 66% year-over-year, with about $370 million from its GLP-1 weight-loss business, according to Sacra's estimates. The company was last valued at $7 billion and has treated millions of patients since 2017, reaching one in every U.S. county and 98% of primary-care deserts. CEO and co-founder Zachariah Reitano has positioned Ro as a distribution partner to pharmaceutical manufacturers rather than a disruptor.

Why do GLP-1 patients stop treatment, and how does intake affect it?#

GLP-1 patients most often stop treatment because of program cost (38.7%), dissatisfaction with results (16.8%), and intolerable side effects (15.4%), and up to 57% discontinue within six months without behavioral support, according to a peer-reviewed telemedicine analysis. Intake affects this because a static form can't surface a patient's prior-attempt history, cost sensitivity, or side-effect fears — the exact signals that would let a care team intervene before a patient quits.

What is AI patient intake and how is it different from a form?#

AI patient intake is a conversational interview, conducted by an AI agent, that collects every clinically required field while also asking natural-language follow-ups to capture a patient's reasoning, history, and "why now." It differs from a form because a form is a fixed decision tree that records answers, whereas a conversation follows the patient — probing vague responses and surfacing context a dropdown can't hold. This makes intake a source of continuous adherence and voice-of-patient signal rather than a one-time gate.

Conclusion#

Ro's AI strategy is genuinely impressive where it has been deployed: a vertically integrated $7-billion telehealth pharmacy using LLMs to triage adverse events with 94% accuracy and deliver clinical outcomes — 16.6% average body-weight loss over 68 weeks — that match the trials. The gap is the front door. Patient intake is the one surface every patient touches, the place where adherence is won or lost, and the place still governed by a static questionnaire that can't ask "why now?" When up to 57% of GLP-1 patients discontinue within six months, the cost of missing that context is not abstract.

Conversational AI intake closes the gap — interviewing patients in their own words, probing the hesitation and history a form flattens away, and turning the intake moment into a continuous voice-of-patient program. That is exactly what Perspective AI is built for. Start a research study or explore intelligent intake to see how conversational intake captures the reasoning behind every patient's first answer — the reasoning that determines whether they stay.