Introducing Midi Prism: The Clinical Knowledge Platform Built for Women’s Health

A Midi visit covers a lot of ground. A woman might come in for hot flashes and, in the same conversation, bring up poor sleep, brain fog, and a family history of heart disease. Her clinician has to connect those dots in real time, with guidance that fits her, not a generic adult patient.

That’s harder than it sounds. Women in their 30s through 60s have been overlooked in medical research for decades, and general-purpose medical tools inherit that gap. At Midi, we’ve spent years turning clinical experience specific to women into protocols, based on guidance from experts in-house and advising specialist physicians, the latest research, and knowledge gained from treating more than half-a-million women, which has led to one of the largest datasets on women between 35 and 65. Our new AI tool, Midi Prism, is how we put that knowledge to work in the flow of care, and enable our clinicians to treat an increasing number of overlapping issues that previously might have required six different visits to six different specialists. 

What Prism is

Prism is Midi’s AI-powered clinical knowledge platform. A Midi clinician can ask a question in plain language — something like “What are the contraindications for hormone therapy?” — and get an answer grounded in Midi’s clinical protocols, broader women’s health data, and peer-reviewed research from PubMed. Each response also shows clinicians the source so they can check the basis.

What matters here is that Prism isn’t a generic chatbot. Prism is built on a library of thousands of expert-built protocols developed from years of Midi’s clinical experience, covering topics from abnormal uterine bleeding and bone health to contraception, cholesterol, and sleep.

In one tool, Midi clinicians get an interactive way to learn the practice’s clinical protocols instead of hunting through documents, and fast, protocol-aligned guidance at the moment they need it; plus, it enables them to quickly and easily check care against guidelines, keeping care consistent as we scale to serve more than 35,000 women a week across all 50 states, and grow our clinical workforce to meet that demand.

One clinician, many specialties

Our bet is that, with the right support, one clinician can deliver connected, expert-informed care across primary and specialty care in a single visit. Prism gives Midi clinicians real-time access to specialized knowledge across gynecology, endocrinology, cardiology, nutrition, sleep, weight management, and more, that helps clinicians address myriad related concerns in a single visit. And, when a specialist referral is the right next step, they’ll recognize that, too.

Physicians and clinical leadership draft, review, and vet the protocols that Prism draws on, and Prism makes them available to clinicians at the moment they need them.* Usage insights show us where a protocol doesn’t yet have a piece of information a clinician needs, and that feedback can then go straight to our protocol committee to inform what we develop next. 

Built from scratch on our own foundation

Prism started on a third-party platform. As more clinicians came to rely on it, we wanted two things: performance that holds up no matter how many clinicians are using the tool at once, and ownership of what makes Midi’s care distinctive. So we rebuilt it on Midi’s own technology stack.

As a result, Prism now holds up during peak demand, allows us to implement a rigorous process for testing accuracy, and gives us rich insights on usage. We fully own the evaluation pipeline, which is how we test accuracy before changes or new information reaches clinicians. And Midi’s protocols and patient data stay secure inside systems we control. It’s also faster, with responses generated in less than six seconds and completed in under twenty. 

With well integrated telemetry, we get rich insights into how clinicians use Prism. About two-thirds of questions are asked right around a patient visit, either just before, during, or between appointments. Clinicians are turning to Prism in the moment they need it, not after the fact. We can also see which types of care prompt the most questions, which shows us where to improve protocols and training. 

We recently brought Prism directly into MidiPal, our AI-powered clinical workspace, so clinicians can use it without switching tools.

As with all AI at Midi, the clinician remains in the driver’s seat. We designed Prism to support human clinical judgment, not to replace it. 

• The clinician makes the care decision. Prism informs it.
• Answers draw on Midi’s protocols and peer-reviewed research, and each answer cits its source so clinicians can verify it
• Clinician feedback feeds back into how we improve the protocols and the tool.

Where we’re headed

Prism is still early, and what we build next will be shaped by what clinicians tell us and what we learn from how they use it. Right now, we’re focused on:

• More context: giving clinicians relevant patient information so answers are specific to the woman in front of them, not just to a condition, so they can create even more personalized care plans
• More knowledge: bringing additional trusted sources into Prism alongside Midi’s protocols and peer-reviewed research
• Better quality, easier to use: improving accuracy, speed, and usability, guided by insights into how clinicians use Prism today

Further out, we see Prism becoming a true co-pilot. Today, a clinician asks a question and Prism supports them with an answer and the source for verification. Over time, we want that support to show up at the right moment in a visit, without the clinician having to stop and type. Think of a knowledgeable colleague who speaks up when it helps. The clinician stays in charge, and the care decisions stay theirs.

Prism is one piece of a bigger effort to give clinicians a single, connected workspace, with MidiPal at the center. We’re excited to share more as we test and build it.

Solving the care gap by marrying expert care with intentional tech

We don’t have enough primary care or specialty providers to treat all the women in this country, and training more of them won’t close the growing care gap on its own. We think technology can help. When a clinician has a field’s worth of protocols and decision support behind them, specific to the population they’re serving, they can cover more ground with more confidence, embodying the AI-enabled “generalist-specialist” role that could help solve one of healthcare’s biggest and most persistent issues: access. It’s what lets Midi offer more kinds of care to women at more stages of life without ever lowering the bar.

We’re very excited to be designing “clinician-in-the-loop” AI with purpose to enable our clinicians to spend more time on care, and give our patients a home base for their healthcare, from age 18 to 80, wherever she may be.

* Clinical services are provided by independent, affiliated medical practices.