Medicaid + AI: A New Standard for Innovation

For too long, Medicaid has been left behind by technology. Innovation has often started at the top of the market and trickled down years later, if it ever arrives at all. Continued advancements in AI and machine learning in healthcare—particularly large language models and agentic AI—are a chance to change that narrative.

We’re in a moment of both urgency and possibility. This is a generational opportunity to set a new standard for healthcare innovation. 

Across the country, care delivery is straining under mounting pressure: policy changes are muddling coverage and eligibility, costs of care are rising, and workforce shortages continue to plague the industry. At the same time, enthusiasm for AI has reached a fever pitch. Every week brings a new breakthrough, a record-breaking valuation, a renewed promise that technology will change everything. Yet very little of that innovation has touched the people and providers who need it most.

As Medicaid programs face shrinking budgets and growing demand, AI offers a way forward.

The secular headwinds we face—economic tightening, administrative complexity, and workforce attrition—are forcing the healthcare system to evolve. They’re also creating the conditions for technology to finally be used in service of scale, quality, and access, rather than efficiency alone.

The content of this report explores what it takes to turn that vision into reality. We outline the principles that should guide AI development, the opportunity AI-driven healthcare applications present in Medicaid, and how Cityblock is using AI, machine learning, and predictive analytics to make our care teams more present, more informed, and more connected to the people we serve.

If we as leaders and innovators can get this right, Medicaid can be the proving ground for effective and responsible AI in healthcare. We can show that the latest and greatest technologies can serve the most vulnerable among us—that the power of AI doesn’t have to accrue with the few. We can build a healthcare system that is made more efficient by being more human and compassionate.

That’s the standard we should set. At Cityblock, it’s the one we strive to meet.

Medicaid stands at a critical inflection point. State programs are being asked to do more with less: funding is constrained, millions are at risk of losing coverage, and providers are struggling to manage larger, more complex populations than ever before. 

Traditional models can’t keep pace with the current environment. Every hour a clinician spends on paperwork is one less with a patient. Missed outreach or delayed follow-up can quickly escalate into otherwise avoidable emergency care. The chasm between medical care, social needs, and operational capacity continues to widen. The system is being pushed to its limits.

Alongside this pressure lies opportunity. As adoption accelerates, AI is transforming healthcare. But that transformation has been uneven. Most investment has focused on administration optimization and revenue capture, not care delivery or member engagement. Medicaid has largely been left out of the conversation altogether. 

But it’s within Medicaid that AI can make the most meaningful, material differences in health and engagement. 

With agentic AI, the healthcare system no longer has to close at 5 p.m. A member can text for guidance at midnight and schedule a visit. A care coordinator can get real-time alerts on emerging risks and intervene earlier. These tools can remake Medicaid, but they must be developed and designed for its realities. That means systems trained on representative data, grounded in empathy, and guided by the principles of responsible innovation.

In a time of economic pressure and rising care complexity, the promise of AI in Medicaid is sustainability. Generating efficiencies is important, but the real opportunity is in making healthcare more accessible, more responsive, and ultimately, more human.

Cost containment is the top priority for most plans:

Building AI that works in Medicaid is hard. The stakes are high, the data is fragmented and deeply contextual, and the margin for error is slim. But the potential impact is enormous.

The healthcare industry has poured extraordinary time, energy, and capital into automating the “easiest” parts of care, i.e. the administrative and financial back-end. Roughly 60% of healthcare AI investment is going toward revenue cycle management, billing optimization, and risk adjustment. These are tools designed to help organizations get paid faster for the work they’ve already done.

It’s a natural starting point, but we can do so much more.

When innovation is concentrated in the revenue cycle, we risk deepening the inflationary spiral that drives up costs without improving outcomes. Risk adjustment, coding automation, and margin expansion will never solve the core problem: how to deliver more care with fewer resources, and build trust with the people who need it most.

Medicaid sits at the center of that challenge and opportunity. Getting this right requires more than powerful algorithms or large datasets. It requires a new framework that puts the people who rely on Medicaid and the care teams who serve them at its center.

At Cityblock, we believe responsible AI in Medicaid must be guided by six principles.

For decades, healthcare innovation has moved from the top down. Breakthroughs designed for commercially insured populations eventually reach Medicaid years later—if they ever do at all. Advanced AI-driven healthcare applications like agentic systems are an opportunity to reverse that trend.

Agentic AI in Medicaid can initiate workflows, coordinate and schedule care, and communicate directly with members, all while keeping human teams in the loop. For healthcare organizations, that means closing gaps between touchpoints in care, triggering follow-ups in real time, and supporting members around the clock, not just within the traditional hours of operation. 

Imagine a care partner in your pocket—a well-informed guide that can help you navigate your care, answer questions about your care plan or medication regimens, and connect you directly to food and housing resources at the moment of need. 

But these models are only as good as the data they’re trained on. Cityblock’s advantage lies in our years of experience working side by side with care teams and Medicaid members, and the deep understanding we’ve cultivated of the members we engage with every day.

We know that with the right incentives and the right technology, AI can direct value where it matters most: those who need care.

Below is what that journey looks like when AI is designed to serve Medicaid members.

We’re living through another technology gold rush. 

The AI boom feels eerily similar to the dot-com era; each promising transformation, both generating enormous value for a select few. We have an obligation to chart a different course this time.

AI has the potential to change not only how healthcare is delivered, but to whom it delivers value. Too often, those who stand to benefit most from innovation are the very last to feel its effects. If we allow this moment to unfold like the tech booms before it, we’ll miss a real generational opportunity to transform our healthcare system for the better.

At Cityblock, we’re proving it’s possible to harness best-in-class AI to serve the most vulnerable among us. That begins with centering equity, trust, and human connection in every system we build. These are the very same principles that have guided our work since our founding.

Every technological revolution is a moral-defining moment. Times like these force us to decide who we want to be as a society and what we believe progress should look like. In healthcare, that means asking ourselves: Do we want to use AI to expand profit? Or do we want to use it to measurably improve lives?

I believe we can build a future where AI amplifies clinical compassion, and where the most powerful algorithms are designed to support people living with the greatest need. That’s a tech revolution worth leading.