Hims & Hers launched an artificial intelligence agent embedded in its platform to help interpret biomarker lab results and provide users personalized insights about their health.
The company launched its direct-to-consumer lab testing program for health biomarker testing back in November. The new agent AI, Labs AI, has been available to some customers in beta testing and will roll out to all Labs customers over time, the company announced Thursday.
Hims & Hers' Labs offers access to 130 biomarker tests across 10 health areas, including heart health, metabolism, hormones, inflammation and stress, as part of its strategy to extend into prevention and health screening. The new AI care agent makes customers’ lab results clearer, more useful and easier to engage with, according to Patrick Carroll, M.D., Hims & Hers chief medical officer.
"The AI companion will educate customers on every lab test that they get back. It is not set up to diagnose," Carroll said in an exclusive interview about Labs AI. "It's set up to format the information and educate customers and that same information that comes from the AI application goes to our provider so they can match up the messaging, and it gives much cleaner, more interpretable results to our customers using AI."
The AI agent can identify patterns across 130 biomarker tests, explain to users why they matter and how they fit into a broader picture of health, executives said.
"It's where AI and healthcare will be going. It’s using the power of an AI tool, but also with the strict oversight of real people, of providers who are actually providing information and background on the AI messaging. The providers get to actually proofread every message that goes out,” Carroll said.
Labs AI is a custom-built AI system that reasons across customers’ biomarker histories and personal health profiles to produce insights that are specific, safe and actionable, according to the company. It combines frontier language models with Hims & Hers' in-house clinical knowledge base and anonymized learnings from thousands of patients.
Labs AI does not pull from the open internet for answers and consumer data is kept secure, Mo Elshenawy, Hims & Hers chief technology officer, told Fierce Healthcare in an email interview.Â
The AI agent does not diagnose or try to replace clinical judgment, he said.
"Labs AI uses frontier foundation models, and grounds responses in customers’ biomarker data and our curated medical knowledge base. That knowledge base is proprietary and was developed with healthcare practitioners specifically for this purpose," Elshenawy said.
He asserts that Labs AI is built with "clinician-approved guardrails, monitored continuously for accuracy and safety, and evaluated against rubrics set by medical experts."
"It includes mandatory cross-category safety checks, tool-grounded retrieval of the patient's biomarker values, and a layered evaluation system that gates every code change," he said.
Elshenawy joined the company a year ago, bringing his expertise in tech and AI, including most recently serving as president and CTO at Cruise, a self-driving vehicle company owned by General Motors.
"What drew me to Hims & Hers was a specific combination I hadn't seen before: a trusted platform that millions of people already use to access care, a seamless model that lets customers actually control the experience end-to-end, and a real need for a modern technology foundation. Most companies have one of those. Very few have all three," he told Fierce Healthcare.
Elshenawy said he sees the potential to use technology to improve the healthcare system, which is still "too fragmented, too reactive, too hard to navigate."
"I joined because I believe we can build something fundamentally better here: an AI-native platform where intelligence is embedded across the care journey—not layered on top of broken workflows — with clinical guardrails, trust at the center, and real outcomes as the measure of success. That combination of mission, foundation, and technical ambition is what made this an easy yes for me," he said.
Hims & Hers executives said the company plans to roll out additional AI care agents.
"The same tool-grounded reasoning, layered evaluation, and tracing infrastructure that powers Labs AI is the foundation for everything that comes next. This is not a one-off feature—it is a platform for embedding intelligence across the care experience," Elshenawy said.Â
"Some of what comes next will be subtle: faster, more seamless experiences that make care easier without calling attention to themselves. Other capabilities will be more visible, like smarter intakes that adapt to what a customer shares, and care companions that help people build and sustain long-term health habits," he said.
Deeper dive into Labs AI
For each customer, Labs AI assembles a structured profile that includes current and historical biomarker values, trends over time, demographic and lifestyle context, and prior care notes, if a customer chooses to share them. The AI agent can surface patterns that may not be obvious from a single biomarker, like a testosterone value that reads differently in the context of sleep and stress results, or a cluster of results that suggest a broad metabolic trend that no single biomarker can show, Elshenawy wrote in a blog post.
Customers can ask the AI agent questions about their biomarkers, health categories and analytics directly. For a customer asking about heart health, Labs AI could explain how a pattern in their LDL cholesterol, ApoB, and lipoprotein(a) biomarkers indicates additional cardiovascular risk. Labs AI can also generate personalized guidance on how aerobic activity, soluble fiber and healthier fats could reduce that risk, as well as when to involve a licensed clinician, according to the blog post.
When a customer needs more than education, the AI agent proactively recommends they connect with a licensed provider. When that handoff happens, the provider already has Labs AI’s summary analysis, so there's no lost context, according to the company.
"What makes Labs AI different is that clinicians were involved from the start, not brought in at the end to review a finished product," Elshenawy said.
The protocols and guardrails Labs AI operates within were built with healthcare practitioners, he noted. "Clinicians also authored the rubrics we use to evaluate accuracy, consistency, safety, and groundedness. Those rubrics are encoded as automated tests in our development pipeline, so every code change is checked against them before it can ship," he added.
Elshenawy's tech teams also regression-test model and protocol changes against benchmarks set by medical experts. "We run dedicated adversarial testing, including prompt-injection sweeps, to make sure the system stays within its intended scope and safety boundaries. The reasoning is simple: if clinicians do not trust the system, it should not be in customers’ hands. Labs AI is designed to educate, not diagnose, and licensed providers remain accessible when customers need clinical care," he said.
Consumers are increasingly turning to publicly available chatbots like ChatGPT to get more health information. But Carroll noted that these AI platforms do not connect to providers.
“We have the advantage that our AI tool is connected directly to a provider group, and those guidelines are reviewed and built out from a real provider group. We're just using the AI tool as a way to format, inform and educate customers. I think it brings great value to our customers, because it's not a one-off AI tool that's actually not woven in or not part of a provider group, and that's a key differentiation for us,” he said.
Many AI tools in healthcare are built for general purposes and then applied to a clinical context. Elshenawy asserts Labs AI was engineered "to use diverse data to surface complex and nuanced patterns related to customers’ biomarkers."
"I would describe the technical approach as retrieval-augmented reasoning, tool-grounded access to the customer’s structured health profile, and a layered evaluation framework that tests every release against accuracy, safety, and groundedness rubrics," Elshenawy told Fierce Healthcare.
"When a customer receives new lab results, Labs AI does not evaluate them in isolation. It brings together current results, historical values, trends over time, demographic context, lifestyle factors, and prior care notes when a customer chooses to share them. From there, it helps explain what has changed, what may be connected, and which combinations of results may be worth discussing with a licensed provider," he said.
The clinical context matters, he added, as two people can have similar biomarker results, but very different histories, risk factors, goals, or patterns over time.Â
"One concrete example is the ability to look across categories of biomarkers, rather than treating each result as a standalone number. A single value may not tell the full story, but patterns across metabolic health, hormones, inflammation, nutrition, and organ function can help customers better understand their whole-body health and where follow-up may be useful," he said.
Carroll sees uses for AI agents beyond Hims & Hers' Labs program, particularly with its weight management program.
“We have built out a first-level AI for GLP-1 communication, and we're building out other levels of that. Today, we use a large group of RNs as care coaches for that high-touch communication. But we're also bringing in AI to support those care coaches,” he said.
“We'll look at every vertical, except for mental health, to see if there's actually an application of AI that can improve the customer experience and can enhance the quality of care,” he added.