In the wake of the 2021 pandemic, former Uber and Handy executive Carl Madi recognized a massive disconnect in the labor market. Thousands of workers were losing service jobs just as the healthcare industry was hitting a breaking point. To help bridge this gap, Madi founded Stepful. By leveraging AI-based simulations and a "flipped classroom" approach, Stepful has streamlined the path to becoming a medical professional, reducing a two-year certification process to just five months. The platform has rapidly evolved into a critical workforce engine, recently securing a $55 million Series C to scale its employer-embedded model.
As CEO, Madi has steered Stepful from a small direct-to-consumer startup into a major institutional partner for some of the nation's largest healthcare systems, including Providence and Johns Hopkins. Under his leadership, the company has scaled from graduating 40 students in its first year to over 22,000 in 2025 – all while maintaining a primary focus on social mobility, with over 80% of enrollees coming from underserved backgrounds. In this interview, Madi details how Stepful is utilizing AI to solve the projected shortage of millions of healthcare workers and why the future of vocational training lies in a decentralized, skill-based model.
Q: You started the company five years ago, back in early 2021, when the pandemic was hitting. Was that part of the driver?
A: I used to work for Uber and then Handy, a company that helped connect households with cleaners. When the pandemic hit, a lot of those cleaners lost their jobs because people didn't want anyone in their homes. I started asking, "What’s a good pathway for these workers to access quality jobs with career progression and growth?"
Healthcare was a clear destination, but when I looked at the pathways to a career or education, they were mostly trade schools or community colleges. That meant it would take people up to two years and cost nearly $20,000 to access an entry-level healthcare position. Worse, the educational institutions were far removed from the hospital and health systems that were employing the graduates. I wondered, how can a cleaner afford that and overcome those barriers? How can they make a living for two years while paying that kind of money? So, we created Stepful to build a better way.
Q: It’s a fascinating model. How are you able to do it so much more affordably and in so much less time than with traditional methods? Is the outcome still the same in terms of credentials?
A: One thing to understand is that learning and the educational system has not evolved in 30 years. It’s still largely lecture-based. When the pandemic hit, everyone just moved that same model onto Zoom. At Stepful, we challenged the traditional model and instead are using an "active learning" experience, which means we flip the classroom. Most of the work is asynchronous, with students learning through bite-sized videos and other content. Most of the synchronous experience consists in Q&A sessions, group projects, or simulations with AI avatars.
AI gives us the opportunity to do things we couldn’t do before without a human instructor. For example, our students need to learn soft skills and how to talk to patients. In the past, they needed an instructor to simulate that. Today, they do it with an AI avatar. At the end, the AI "patient" becomes an AI "preceptor" and gives them immediate feedback. It makes it much more engaging because students don't have to wait a month for an instructor to get back to them - all feedback is real-time.
Q: When did you start using this avatar? Does it fully replace humans in the practical feedback portion? And, how else do you use AI in your curriculum?
A: About two years ago. It doesn’t fully replace humans, but it is used in about 90% of the student interactions. We still like to have a human-in-the-loop for certain checks. As we move away from "number of hours" toward "competency-based" learning, we will keep iterating to find the quickest way for people to gain and evaluate all the required skills, at scale.
We also use AI to evaluate competencies. For instance, we train people to do blood draws. We send them a clinical kit with a fake arm, and the student performs the procedure in front of a camera. An AI instructor walks them through the procedure, step-by-step, and gives them feedback.
Q: When they are actually tested for certification, is it a third party doing the testing?
A: It depends. All of our programs lead to a national certification or licensing exam. Most of these exams are knowledge-based, meaning they don't require showing practice, just knowledge. They are typically online exams, and our pass rate is very high – usually 5% to 10% above the national average.
Q: Which professions does Stepful help people enter?
A: Our goal is to cover all of allied healthcare and nursing. We started with medical assistants, patient care technicians, phlebotomists, and dental assistants. We’ve since launched more advanced certificates, such as surgical technicians and sterile processing technicians. Now, we are launching degree-based certifications like Registered Nursing, respiratory tech, and imaging tech. Nursing is already live; respiratory and radiology tech are going live in Q1 2027.
Q: You’re moving from a direct-to-consumer (D2C) model to becoming more of an end-to-end workforce solution for hospitals. Are you fully transitioning, or will you still offer D2C? And how are you balancing that high-touch placement with a scalable AI platform?
A: We are definitely going to keep investing in the direct-to-consumer side. There, we’re doubling down on entry-level roles – helping people go from "zero to one" to enter the field. Part of the draw of our D2C platform is that students have access to a career search platform that automatically links them to approximately 40,000 clinical sites. And this high grad placement rate is, in part, what helped get attention from hospitals and health systems.
We started our B2B offering because healthcare systems came to us, impressed by our graduates, and wanted to use our training to develop their own personnel. Today, we work with around 35 healthcare systems, including Mount Sinai, Providence and Ochnser. Our solution allows them to train staff from entry-level roles into advanced roles like Registered Nursing.
Our mission is to break down barriers. D2C makes healthcare jobs affordable and accessible. B2B is a continuation of that, allowing us to offer advanced pathways that the employers actually pay for.
Q: In the vocational space, speed to certification is paramount. How does your five-month timeline change the unit economics for a health system compared to a two-year community college track?
A: It improves it by 4x. A community college track is 4x as long, which means the healthcare system can't benefit from that medical assistant for 2 years. Offering a faster time to practice allows the system to see more patients and generate more revenue much sooner.
Q: You recently introduced "employer-embedded" education. Are you co-designing the curriculum around specific hospital workflows?
A: Most of our programs align with professional certifications approved by third-parties, so there isn’t much customization in the core curriculum. However, we do take employer feedback to improve the programs. More interestingly, as AI becomes mainstream, these jobs are changing. We are currently working on a program to help healthcare professionals use AI in healthcare settings, and we will co-design it with healthcare systems.
Q: Where do you see AI playing a role in the "less sexy" parts of the business, like vetting candidates or predicting dropouts?
A: Most online programs have pretty poor graduation rates – between 10% and 20%. Community colleges are around 50%. Our graduation rate is above 70% for D2C and above 90% for B2B.
We use AI to predict whether a student will drop out by analyzing their behavior on the platform. Are they missing quizzes? Have they stopped attending live sessions? We can tell if someone is deteriorating from their previous behavior and then generate AI-driven interventions to coach them back toward their goal of completion.
Q: You offer a graduate guarantee, where companies only pay for successful graduates. Do you have data on how successful those AI interventions are in helping people graduate?
A: We run everything based on data experimentation. It’s been a gradual lift, but our completion rates are 2x to 3x higher than non-AI environments. Another factor is employer retention. We’ve conducted case studies showing a 60% to 70% improvement in retention after year one. This is because we do a lot of screening up front, and because upskilling internal employees – like a receptionist becoming a medical assistant – means they already know the culture and the manager.
Q: The AHA is projecting a shortage of 3.2 million allied health workers by the end of this year. What is the key to Stepful’s growth?
A: Healthcare systems typically rely on sign-on bonuses or travel agencies to fill gaps, which costs them $100 billion a year. Those solutions don't create new supply. Trade schools and community colleges are supply-constrained because they are brick-and-mortar-based. They have fixed student-to-teacher ratios and limited lab space.
The second issue is clinical rotations. Community colleges struggle because there is no alignment of incentives; a hospital trains a student for free, and then the student goes to work for the highest bidder. In our B2B model, clinical rotation is not an issue. If a hospital is upskilling its own assistant to become a nurse, it is happy to host the rotation because it knows the worker will stay with it.
Oak is thrilled to back Carl and the Stepful team again, having witnessed firsthand his ability to identify underserved markets and build scalable solutions that create real impact. From graduating just 40 students in its first year to over 22,000 in 2025, Stepful's growth trajectory speaks for itself – and that's before accounting for the massive runway ahead as the healthcare industry faces a projected shortage of 3.2 million allied health workers. With a rapidly expanding roster of institutional partners, a proven employer-embedded model, and AI capabilities that are only getting stronger, Stepful is uniquely positioned to become the defining workforce platform in healthcare.