A public health researcher and an AI entrepreneur have joined forces to tackle one of healthcare's most persistent challenges: helping people find and access the services they actually need. The partnership, formalized this week between Professor Bhavnani at the University of Texas Medical Branch (UTMB) and CEO Kumar of Deep Impact AI, represents a different kind of AI healthcare story, one focused not on diagnosis or drug discovery, but on the unglamorous work of connecting vulnerable populations to existing safety-net resources.

What Problem Does This Partnership Actually Solve?

The collaboration centers on an application called Resource and Access Precision Policy, or RAPP, which Bhavnani's team developed with the Houston Health Department. The tool uses AI to help residents navigate the fragmented landscape of public health services, from emergency care to chronic disease management. Rather than asking patients to figure out which clinics, programs, or benefits they qualify for, RAPP aims to match people with the right resources based on their specific circumstances.

This focus on the "last mile" of healthcare delivery reflects a growing recognition that AI's biggest impact may not come from flashy diagnostic breakthroughs, but from solving the messy, real-world problem of access. As Bhavnani explained, the partnership is grounded in something more personal than pure technology: both leaders have experienced firsthand how chronic illness can devastate families when people cannot navigate healthcare systems effectively.

"Unwritten in our agreement is a shared belief: translation must be grounded in alleviating human suffering informed by our own personal stories," said Prof. Bhavnani.

Prof. Bhavnani, School of Public and Population Health at UTMB

How Will This AI Tool Actually Work in Practice?

Deep Impact AI will build a production-ready version of RAPP, while Bhavnani's team provides expertise on the underlying algorithm and user interface design informed by feedback from actual patients and healthcare workers. The goal is to create an enterprise-level solution that the Houston Health Department can deploy and eventually scale to other cities.

The partnership also reflects a broader shift in how AI is being integrated into healthcare. Rather than replacing clinicians or making autonomous decisions about treatment, tools like RAPP are designed to make existing systems more efficient and accessible. This approach sidesteps some of the thorniest challenges that have plagued other AI healthcare initiatives, including questions about liability, accuracy, and whether algorithms can truly be trusted with life-or-death decisions.

Steps to Translate AI Research Into Real-World Healthcare Solutions

• Deep Stakeholder Engagement: Both the researcher and the company worked closely with the Houston Health Department and community members to understand the actual workflows and pain points before building the tool, ensuring the AI solution addresses real problems rather than theoretical ones.
• Clear Intellectual Property Agreements: The partnership required careful legal work to define ownership of jointly developed software, a step that many academic-industry collaborations overlook but which is essential for sustainable long-term partnerships.
• Balancing Functionality With Feasibility: The development strategy prioritizes what is actually achievable given real-world constraints, including limited resources and a dynamic policy environment, rather than pursuing a perfect but impractical solution.

The Bhavnani-Kumar agreement also highlights a challenge that often goes unmentioned in AI healthcare discussions: the difficulty of moving from a successful research prototype to a tool that actually works in hospitals, clinics, and public health departments. Legal teams from both UTMB and Deep Impact AI spent considerable time clarifying ownership and governance structures, a process that underscores how translation involves far more than just engineering.

Why Does Medical Education Need to Adapt to AI Now?

While the RAPP partnership focuses on public health infrastructure, the broader healthcare system is grappling with a parallel challenge: how to train the next generation of physicians in a world where AI tools are becoming routine. Dr. Jennifer Nelson, program director of the Internal Medicine Residency at the University of Arizona College of Medicine, has been thinking deeply about this question.

Nelson notes that residents are already encountering AI in clinical practice through ambient listening tools that automatically transcribe patient conversations, AI-powered clinical scribes, and diagnostic support systems. The challenge for medical education is integrating these tools in ways that enhance physician training rather than shortcut it.

"The question for GME is how we integrate these tools into training in a way that helps our residents become better physicians without allowing technology to replace the foundational skills they need to develop," explained Dr. Nelson.

Dr. Nelson, Program Director of Internal Medicine Residency, University of Arizona College of Medicine

Nelson envisions a future where electronic health records incorporate tools that identify personalized learning opportunities for individual residents, moving away from one-size-fits-all curricula toward adaptive education that responds to each trainee's needs. This approach mirrors what the RAPP partnership is attempting in public health: using AI not to replace human judgment, but to make human expertise more efficient and accessible.

How Is AI Changing Dementia Research and Care?

Beyond public health and medical education, AI is also reshaping how researchers evaluate and deploy new technologies for serious conditions. UCLA Health recently received a five-year, $25 million grant from the National Institute on Aging to lead a nationwide program for testing and evaluating AI tools designed for patients with Alzheimer's disease and other dementias.

The initiative, called Accelerating Platforms for Evaluating and Executing AI Lifecycles in Alzheimer's Disease and Related Dementias (APEX-ADRD), involves collaboration with Mayo Clinic and the University of Wisconsin. With over 7.2 million Americans living with Alzheimer's disease, a number expected to rise in coming decades, the program aims to help clinicians identify cognitive decline earlier, personalize care plans, and reduce administrative burden.

What distinguishes APEX-ADRD is its commitment to rigorous evaluation across diverse populations. Rather than testing AI tools in a single health system or demographic group, the program will validate findings across California's diverse population and partner institutions, ensuring that AI benefits work for everyone, not just those who happen to be well-represented in training data.

"AI has enormous potential to improve dementia care, but potential is not the same as evidence. The ultimate measure of an AI tool is not how impressive the technology is, but whether it improves care for patients and families," stated Dr. Joann Elmore.

Dr. Elmore, Professor of Medicine at the David Geffen School of Medicine at UCLA

The UCLA initiative reflects a maturing approach to AI in healthcare: rather than asking whether AI can do something impressive in a lab, researchers are asking whether it actually helps real patients in real clinical settings. This shift from "can we build it?" to "does it work for everyone?" represents a fundamental change in how the healthcare system is approaching artificial intelligence.

These three developments, spanning public health infrastructure, medical education, and dementia research, suggest that AI's most meaningful impact on healthcare may come not from dramatic breakthroughs, but from the harder work of making existing systems more equitable, more efficient, and more responsive to human needs.