Most decisions about which AI tools enter a hospital are made far from the bedside — in boardrooms, procurement offices, and vendor pitch meetings. In June 2026, the American Nurses Foundation tried a different approach: it announced the inaugural awardees of its “Nurse-Led Solutions Using AI and Innovation” microgrant program, handing three nurse-led research teams $10,000 each, funded by the AI healthcare agent company Hippocratic AI, to study AI’s real-world effects from the inside.
Putting Money in Nurses’ Hands, Not Just Executives’
The microgrant program inverts the usual order of operations in health-tech adoption, where administrators typically select and deploy a tool before frontline staff have much say. Applications for the program were due February 17, 2026, and winners were notified that spring. Each of the three funded teams is evaluating a different facet of AI’s impact: patient safety, preventable deaths, and the completeness and accuracy of clinical documentation. The grants are structured to support the full arc of a research project — design, implementation, data collection, tools and resources, and dissemination of findings — rather than a one-time survey or white paper.
Why Hippocratic AI Is Funding Nurses to Scrutinize AI
The funding comes from Hippocratic AI, a company building generative-AI “agents” for non-diagnostic clinical work: nurse phone calls, patient check-ins, and chronic-disease coaching calls that are typically labor-intensive and repetitive. Hippocratic AI has raised more than $400 million across its funding history, including a $141 million Series B that pushed it to unicorn status, and a $126 million round in November 2025. Funding independent nurse-led research into AI’s effects is an unusual move for a company whose own products sit squarely in that research territory — a bet, evidently, that credible third-party scrutiny from working nurses will build more trust than marketing alone.
The Backdrop: A Nursing Workforce Under Strain
The program lands amid a nursing shortage that has strained hospitals for years, with burnout and staffing gaps pushing health systems toward automation for tasks that once required a human voice on the phone. That same shortage is part of what makes AI phone agents and documentation tools attractive to hospital administrators — they promise to offload some of the routine calls and paperwork burden from an already stretched workforce. But it is also precisely what makes frontline evaluation so important: nurses, not software vendors, are the ones who will notice if an AI system creates new safety gaps rather than closing old ones.
Skepticism From the Floor
Not every nurse greets AI-agent tools with enthusiasm. Concerns persist that automated phone check-ins and AI-driven chronic-disease coaching could miss subtle verbal or emotional cues a human caller would catch — the kind of judgment call that underpins nursing practice. Documentation accuracy is its own flashpoint: critics of AI scribing and charting tools have warned that automation can quietly propagate errors into a patient’s record if unsupervised. That’s precisely the kind of question one of the three funded teams is investigating directly, rather than taking a vendor’s claims at face value.
A Grassroots Model for Building Trust
The microgrant structure — small dollar amounts, but real ownership of study design in nurses’ hands — offers a template that could scale well beyond three teams if it proves useful. Rather than a single large study commissioned by a hospital system or a vendor, dispersed nurse-led research generates independent data points from different clinical settings, which can be harder for any one stakeholder to spin. Foundations and companies watching adoption battles play out elsewhere in health tech may see appeal in a lower-cost, higher-trust model.
What’s Next
The real test comes when these three teams publish results later in their project timelines. Findings that show measurable gains in documentation accuracy or safety outcomes would give Hippocratic AI and similar companies concrete, independently generated evidence to point to — something the AI-healthcare sector has generally lacked. Findings that surface new risks would carry credibility precisely because they came from working nurses rather than a company’s own marketing materials. Either way, the American Nurses Foundation has set a precedent: evaluating AI in nursing doesn’t have to start and end in a hospital’s IT department.
