Six Opportunities for Digital Health in 2025

1. Co-design is the key to M&A survival‍

Many digital health ventures that raised post-pandemic bridge rounds or early AI funding are running out of capital, which will trigger a new wave of mergers and acquisitions. The companies that survive won’t be the ones with the best tech or the most funding. They’ll be the ones that have built deep, system-wide relationships with health systems and payers.

Too many startups fail because they design for only one stakeholder—patients, providers, or payers. But viability only happens when all three are aligned. And the only way to achieve that is by designing at the intersection of desirability (what people want and need), feasibility (what is technically and operationally possible), and viability (what is financially and systemically sustainable). The solutions that last won’t just check one of these boxes—they’ll be deeply integrated into the way healthcare actually works. And to do that, the companies behind them will need to co-design with all three categories of stakeholders.

That’s what we did in 2024, when we partnered with the California Department of Mental Health and Kooth to create and launch Soluna, a free mental health and coaching platform for youth across California. To create a resource young people would actually use, we worked with more than 150 youth between 13 and 25 across the state to understand their needs, habits, and wants, shaping a platform that felt accessible, engaging, and supportive. To ensure Soluna could scale and actually work within the behavioral health system, we also navigated the constraints of government needs and co-designed with mental health practitioners to create back-office tools adapted to existing workflows. This is what co-design looks like—not just designing what’s desirable for users, but making sure it actually fits into existing systems.

2. Trust is crucial for high-stakes AI products

Some 70 percent of payers and providers have invested in AI to help with documentation, scribing, and automatable admin work. But the true potential lies in patient-facing care, like diagnostics, care delivery, and even clinical trials—but only if we design AI tools that providers and patients actually trust. This year, AI will  increasingly move into at-home care, especially in metabolic health and mental health. With GLP-1 drugs and metabolic surgeries on the rise, patients are going to need long-term personalized behavioral support. AI is perfect for that. In mental health, AI could help close the massive access gap by offering real-time support where no providers are available.

But none of this will happen unless AI is designed with patient trust in mind. We’ve seen firsthand how hesitant health plans are to extend AI beyond internal operations. But AI in healthcare isn’t just a data problem—it’s a human problem. The companies that figure out how to make AI feel safe, clear, and integrated into care delivery will define the next era of digital health.

3. Point solutions are great—but human-centered integration will make them more effective

Digital health is full of point solutions—condition-specific digital therapeutics like in diabetes, sleep issues, and brain health. These solutions have been critical in proving that traditional models of care can be improved with technology. But now, patients are juggling a phone-full of apps; a patient who downloads an app for meditation might need another for virtual primary care, another for weight management, and still another for chronic stress, each siloed in care team, data, and clinical recommendations.

The real opportunity lies in figuring out how to integrate these point solutions into conservative care models, trusted patient-facing brand names, and existing technology/EHR workflows. That means linking them to primary care, specialty care, and the broader health system—not expecting people to manage their own fragmented ecosystem of tools.

‍Maven Clinic does this well. Its mental health offering isn’t just a standalone app; it’s a wraparound service that connects to broader care. That’s the kind of design work we are hoping to see across digital health. As funding tightens and patient adoption slows, the future of digital health will be defined by integration.

4. AI-driven tools will soon replace entire departments

Digital health companies aren’t just pitching software anymore; they’re pitching end-to-end replacements for what were once human-run functions. Startups like InpharmD are using AI to deliver pharmaceutical recommendations at the clinic level, while Hippocratic AI is exploring drop-in AI nursing roles. These aren’t just AI-powered tools—they’re AI-powered workforces.

This isn’t just a shift in technology; it’s a shift in how healthcare organizations operate. If AI agents are going to supplant entire human functions, we need to think beyond just efficiency and automation. What happens to oversight? To decision-making? To the patient experience? The challenge is no longer just about designing better workflows. It’s about reimagining the future of work in healthcare.

5. New consumer choice offers a design opportunity

As people lose trust in the system, they’re looking for ways to take control of their healthcare choices. The question is: How will they actually make them?

With pricing transparency, new payment models where employees buy their own insurance and get reimbursed, potential regulation of pharmacy benefit managers, and a growing dissatisfaction with employer-based insurance, consumers are increasingly getting to choose which provider to see, what treatment to try, and how much they are willing to spend. But the interaction for how that all happens is still up for grabs. Will it happen through provider-led conversations? Employer benefits and brokers? Text-based interfaces like ChatGPT? Or will a new wave of agentic AI help people weigh cost, risk, and treatment options in real time?

So far, no one has cracked the interaction model for healthcare choice. IDEO and Newton-Wellesley Hospital pioneered one method with a concept called “Care Modes,” that allowed patients to decide the approach for how they receive care. The chosen mode flowed into every interaction at the hospital, even adjusting how health information gets logged in the EHR. But options like this are just the beginning.  The future of how people interact with complex healthcare decisions is still wide open. The challenge ahead isn’t just about giving people more data—it’s about designing how they engage with it, trust it, and act on it.

6. Designing for behavior change is still the foundation

Designing for behavior change is one of the most competitive and well-researched areas in healthcare, but there’s still so much room to grow. For the last decade, I’ve taught a course on Designing for Behavior Change at Stanford, where students work directly with patients to develop real lifestyle changes. Over the years, we’ve learned that the most effective interventions involve more than just habit tracking or designing nudges; they also create systems of accountability, community, and self-awareness in a way that aligns with people’s actual lives. That’s why observation and in-context research matter so much.

There’s a big opportunity to rethink how digital health supports long-term lifestyle change—whether that’s reimagining SNAP and WIC programs, integrating AI into personalized coaching, or pulling insights from the most engaging consumer and social platforms.

Prioritizing human-centered design

With so many problems to solve in our health care system (not to mention new ones popping up), there’s no limit to the possibilities for innovation in digital health. But our job as designers is bigger than just making things work; it’s to make a healthcare system that prioritizes user choice, deepens trust in the system, and uses the newest cost-saving technology, while integrating into the existing ways patients access care. That’s how we can drive radical transformation, by putting people—not just business or technology—at the center.

Thanks to Jennifer Lo, Ari Adler, Juho Parviainen, Jesse Fourt for suggestions.