Featured by Newsweek & World Class Media Outlets
Mohamed Elshenawy

Mohamed Elshenawy

Chief Technology Officer
Hims & Hers
03 November 2026

Hims & Hers Health is a San Francisco-based digital healthcare company founded in 2017, providing telehealth consultations, prescription treatments and personalized care through its online platform. Its services span weight management, sexual health, dermatology, mental health and hormonal health, combining virtual care with medication delivery and subscription-based treatment programmes.

You spent much of your career developing AI and autonomous vehicle technology, including leading Cruise, before moving into healthcare. What convinced you that healthcare was where you wanted to apply your experience?

It certainly looks like a sharp pivot, but with advances in deep learning and AI, a lot of knowledge transfers between industries. What changed my thinking was meeting our founder, Andrew Dudum. I hadn't appreciated the scale of the problems in healthcare or the gap between what we should consider acceptable and what people actually experience.

The US spends roughly twice as much on healthcare as other developed countries, yet life expectancy is around four years shorter. To me, that's not fundamentally a biology or science problem; it's a delivery and access problem. Hims & Hers interested me because it has a closed-loop system, from a patient's initial need through treatment to outcomes. That creates an opportunity to apply technology across the entire care journey, rather than improving one isolated part.

Hims & Hers began by addressing often-overlooked health concerns through telehealth. How is the company evolving towards more proactive, personalized healthcare?

Our mission hasn't changed: providing excellent care for everyone, everywhere. But traditionally, someone would come to us for a specific treatment and leave. With the launch of Labs, we're moving towards a more proactive, long-term relationship, supporting people across different aspects of their health over several years.

Someone might receive blood test results covering 130 biomarkers and be left with a confusing report. They could upload it to a general-purpose AI tool and still struggle to understand what to do next. We want to interpret those results in the context of their health history and treatments, explain what they mean and turn that information into action. If something needs clinical attention, we can bring in a provider. It's the difference between giving people more data and helping them use it.

AI health tools are proliferating, but many leave patients responsible for interpreting advice and navigating care themselves. What makes Hims & Hers' approach different?

In traditional healthcare, you might leave an appointment with a prescription, but the physician doesn't necessarily know whether you filled it, started treatment, experienced side effects or stopped taking it. General-purpose AI can answer questions, but patients still have to provide their own medical context and figure out what to do next.

Our system connects those steps, from the initial consultation through treatment adjustments and outcomes. We've introduced a single conversation that brings together the patient, AI, care coaches, and licensed providers. The AI draws on our clinical knowledge and knows when a provider needs to step in. It's not a choice between AI and doctors. We want AI to elevate doctors, giving them better information and allowing them to focus on where their expertise matters most.

Weight management is particularly difficult when it comes to sustaining long-term results. How can AI support patients whose treatment journeys differ so much?

That's precisely why we started with weight management. It's intensely personal, and people's biology, psychology and circumstances vary enormously. One patient might want frequent reassurance, while another needs very little intervention. We wanted a system that doesn't just answer questions around the clock, but reaches out proactively when someone needs support.

It might recognize that a dose adjustment is approaching and help someone prepare for changes in appetite or side effects. Connected scales can reveal when weight loss has plateaued or is progressing faster than expected. If treatment needs adjusting, a licensed provider makes that decision. That's what I mean by operating in the action layer rather than just the advice layer. As we expand into hormonal health, the same approach can support people managing multiple conditions or treatments.

You've compared today's healthcare experience to renting a DVD from Blockbuster. How far are we from delivering the personalization people now expect in other industries?

We've become accustomed to a healthcare experience that hasn't changed in decades. You call for an appointment, wait several weeks, sit in a waiting room and have a brief consultation. I call that the waiting-room standard. What's remarkable is that we're struggling even to meet it because there aren't enough healthcare professionals.

Imagine instead having a dedicated healthcare team that knows your history, understands your needs and is available when you need support. I have close friends who are doctors, and I've seen the difference it makes when someone genuinely knows how to communicate with you. We should be able to make that experience available much more widely. But we can't achieve it simply by adding more doctors. AI can remove routine work so clinicians spend more time on judgment, empathy and meaningful conversations.

Your background in autonomous systems involves complex questions of safety and trust. As AI becomes more involved in healthcare, how do you ensure clinical accountability remains with qualified professionals?

AI doesn't diffuse accountability at all. Our systems aren't there to make clinical judgments that require a licensed provider. The challenge is knowing where that boundary lies and involving a clinician whenever necessary, including when there's uncertainty. Sometimes a human is needed even if AI could answer the question, because someone is distressed and needs reassurance.

We use an approach I call the four Gs: gated, grounded, graded and guardrailed. That means controlling when AI is used, grounding responses in clinical guidance, evaluating performance and keeping it within defined boundaries. We also evaluate AI responses with human oversight and an emphasis on traceability. There's an enormous difference between building an impressive AI demonstration in a few hours and deploying something that works reliably for millions of people. Trust requires engineering safeguards into the system, not adding them afterwards.

You've argued that AI could make healthcare more expensive if it's simply added to existing systems. What would it take for AI to improve access and affordability rather than become another premium service?

The problem comes when organizations bolt AI onto existing services and then look for someone to pay for it. We start with clinical outcomes and redesign the care experience around them. If customers aren't achieving their goals, they won't stay with us.

When we launched AI support for weight-management customers, engagement tripled. Yet AI handled around 80% of those interactions end to end, while the workload passed to human teams was roughly halved. Normally, more engagement means more staff and higher costs. Here, we're seeing the possibility of the opposite: more support for patients while clinicians focus on the interactions that genuinely need them.

It's not about replacing human care, but making more personalized healthcare accessible to far more people.