Why Nurses Need to Build Responsible AI Tools Now, With Sharonda Davis

Sharonda Davis is a PCU nurse at a Level I trauma hospital, a mom of teenagers, and a self-described futurist who sits on the Global Nursing AI Alliance™ (GNAA™). She also just did something most nurses have never considered possible: she built an AI-enabled web platform from scratch, with no coding background, to help other nurses submit public comments on how Medicare should pay for artificial intelligence in clinical care.

On this episode of the Love n’ Leary Nursing Podcast, hosts Marion Leary, PhD, MPH, RN, and Rebecca Love, RN, MSN, FIEL, sit down with Sharonda to talk about the platform she built, what pushed her to build it, and why she thinks every nurse right now needs to pay attention to a few policy conversations that will shape the profession for decades.

The context, briefly: the Centers for Medicare & Medicaid Services released its proposed Calendar Year 2027 Physician Fee Schedule in July 2026, and inside that rule was a Request for Information on how Medicare should value and structure payments for AI in clinical settings. CMS referenced the three-category framework drawn from the American Medical Association’s CPT work on medical AI, which classifies AI services as assistive, augmentative, or autonomous. The public comment window closed on September 14, 2026.

Sharonda built a tool called Nurses Voice Matters to help nurses write and submit comments to CMS before that window closed.

>>Listen to Why Nurses Need to Build Responsible AI Tools Now, With Sharonda Davis

Want to see more Nurse.org articles in your Google results? Add us as a preferred source.

Sharonda is honest about where she started on AI. “I think that I initially approached it very liberally and like yes, this is the future, this is what we need. And this really gave me pause and I wanted to be a part of the moment and contribute to it.” She credits a series of workshops on the CMS process and CPT codes for lighting the fire, saying the material “even caused me to question my stance on AI and healthcare.” That reckoning is a theme nurses listening to this episode will recognize. A lot of nurses are being told AI is coming for parts of their work, that it’s going to be reimbursed, and that it will reshape documentation, workflow, and clinical decision-making. Very few feel equipped to say anything back.

“I think that I initially approached it very liberally and like yes, this is the future, this is what we need. And this really gave me pause and I wanted to be a part of the moment and contribute to it.”

What Sharonda realized, listening to the CMS discussions, was that nursing work is largely invisible in the way these conversations get framed. “Our work as nurses is hidden in the documentation, it’s hidden in conversation, it’s hidden in so many things that when I ask my friends that aren’t in the healthcare space, what do nurses do, the answers vary so widely.” If AI reimbursement is going to be built around the physician’s work, the practice expense, and what gets considered medically necessary, then what nurses actually do at the bedside has to show up in the record somewhere. Public comment is one of the few places it can.

The platform Sharonda built, Nurses Voice Matters, didn’t come from a software background. She can’t code, and she says so plainly on the episode. What she did have was a stack of AI tools and a willingness to try. She took her notes from the workshops she attended, uploaded them to ChatGPT, and asked a very simple question: how could I contribute? From there she used what she calls “a combination of Claude Code, ChatGPT Codex” to write the code, with each one checking the other, and pushed everything to GitHub for a final automated review before deploying.

“I thought about what I would want, and that’s what I encourage all nurses to do in this era, in this rapidly advancing era, is to think about what would help you.”

The result was a prompting flow designed for a nurse who’s never commented on a federal rule before. It walked users through their own clinical skills and knowledge, pulled out what was relevant to the reimbursement debate, and helped them produce a written comment they could submit. By Sharonda’s count, 50 nurses clicked through and 30 ended up filing a comment. She wishes the number were higher, and she’s open about why it wasn’t: she had just started a new job and was cautious about social media policies, so she held back on promoting it. For a first-pass tool built solo by a bedside nurse with no coding background on a hard deadline, those are not bad numbers. They are, as Marion puts it on the episode, a pilot test that proved the concept.

One of the most useful parts of this conversation is Sharonda’s working definition of responsible AI, which is not a slogan but a set of habits. The first is resource awareness. Her teenage sons have been on her case about water consumption from AI model inference, and she takes it seriously. “I think that in some ways I may have been irresponsible because I’ll use like four different tools and I’m just like prompting all of them and reprompting and reprompting.” She has started limiting how many tools she chains together, which is a small discipline that nurses can carry into any clinical AI they end up using.

“Prompting from a problem that you already understand and you already know the outcome because you don’t want to be misled by the AI.”

The second habit is what she calls prompting from a problem rather than prompting from a deficit. In nursing documentation, you only chart when something’s wrong. In AI, she argues, nurses should start from a problem they already understand well enough to catch the model if it drifts. That is one of the clearest frames for safe AI use at the bedside that nurses have offered on this show, and it’s rooted in clinical reasoning rather than computer science.

The third habit is honesty about mistakes. Sharonda mentions on the episode that an earlier project of hers, a nurse prompt hub she built where more than a thousand nurses created prompts, taught her something uncomfortable. Shadow AI tracking can pick up prompts from different contexts and, in theory, reconstruct patient information. “I kind of had to proceed with caution.” Nurses Voice Matters was deliberately designed to be the safest build she could ship. The willingness to say that out loud is unusual, and it matters.

FHIR and the Open Door for Nurses to Build

Sharonda spends a chunk of the episode on FHIR, the Fast Healthcare Interoperability Resources standard, and makes a case that nurses listening should at least know what it is. FHIR is the standard that lets outside tools talk to the big EHR systems, Epic and Oracle Health among them, and it opens the door for nurses to build small applications that solve the small, maddening workflow problems they live with every day. Her example is bedside report. If waiting for report is one of the things that reliably eats your shift, she suggests, that is a candidate for a tool.

“If AI is someday going to be paid for what it’s learning from us, right? It can’t duplicate our gut feeling, but it’s certainly learning from our gut feeling.”

This is the part of the episode nurses might save for later. The build-your-own-tool framing has risks, and both hosts and Sharonda are careful not to pretend it’s easy. Integrating with Epic through FHIR takes institutional access, time, and often a formal application process. But the broader argument that nurses should be building the tools that get used on them, rather than only consuming them, is one of the clearer rallying cries this show has aired.

100 Years of Nurses and the Case for Building Now

The last beat of the episode is personal. Sharonda comes from more than a hundred years of nurses in her family. Almost every woman in her family has worked in healthcare in some form. She spent five years trying to build a startup called NurseBlock, a credentialing and decentralized-care concept that came out of her early work in the Atlanta blockchain community, and she describes that stretch honestly as five years of trying and failing and learning. Coming back to the bedside, she says, has given her a front-row seat to where the gaps really are.

“I come from over a hundred years of nursing. Almost every woman in my family has worked in healthcare in one way or another. So I wanted to come back to my roots and bring all of the things that I’ve learned to who I am and leave that mark here on the world.”

Rebecca and Marion end the episode on an argument that nurses should keep hearing. Nurses who build, build with integrity. If AI in healthcare is going to go in a direction the profession can live with, nurses have to be among the people writing the tools, writing the comments, and writing the next version of the standards. That doesn’t require a degree in computer science. It requires picking a problem that bothers you at work, and starting.

Listen to the full conversation with Sharonda Davis on the Love n’ Leary Nursing Podcast, or wherever you listen to podcasts.

More from Nurse.org’s reimbursement series:

🤔 What’s the one bedside problem you’d build an AI tool to solve if you had the time and the help? Share your thoughts in the comments below.