Recently, my colleague Brian Haenni and I hosted a webinar as part of the Infinitus HAI25 series looking at the trends, adoption, and opportunities across the healthcare landscape. As part of that webinar (now available on-demand here), I made three predictions for transformative shifts coming in the healthcare landscape that are driven by the rapid adoption of AI technology. 

But what we discussed during the webinar isn’t all that I foresee coming in the year ahead. Here are my predictions for healthcare AI in 2025:

1. 2025 will be the year of AI agent networks

2024 was the year of AI scribes in healthcare. 2025 will bring networks of AI agents to the space, thanks to the growing complexity of patient support needs and the rapid advancement of enabling technologies. 

Healthcare workflows often require coordination across multiple domains, from verifying insurance to managing specialty drug authorizations and monitoring patient adherence. Current single-purpose AI agents, while effective, cannot fully address these interconnected challenges. By leveraging networks of AI agents, each specialized in a particular function and capable of seamless collaboration, healthcare organizations can create a dynamic system that adapts to diverse and evolving needs. Advances in AI reasoning, interoperability, and multi-modality are accelerating this shift, making it feasible to implement multi-agent systems that deliver more holistic, efficient, and personalized care solutions.

2. AI phone calls will be everywhere – and you’ll prefer them

In 2025, we can expect AI agents to handle a significant portion of healthcare phone-based communication, both payor-facing and patient-facing.

Why will AI agent calls be preferred to phone calls with human workers? AI is infinitely patient; it won’t lose its cool with busy call center workers at insurance companies if it takes longer than expected to receive an answer. Or, as another example, imagine you’ve been diagnosed with a chronic condition and prescribed a new medication. An AI agent could help you acquire answers to urgent questions that arise at 9 p.m. on a Sunday, when a doctor or case worker likely would not be available. 

In short: 2025 will be the year AI agents become an indispensable part of patient services.

3. AI will redefine roles amid workforce shortages

Another shift I foresee is the role AI will play in addressing workforce shortages within the healthcare sector. These labor shortages are a major deal; the World Health Organization estimates a projected shortfall of 10 million health workers globally by 2030. AI can – and will – be part of the solution.

By integrating seamlessly with human workers, AI will help alleviate the strain on overburdened healthcare professionals. Administrative burdens, redundant tasks, and even some front-office responsibilities can be delegated to AI, allowing medical personnel to focus more on patient care. 

AI will not replace human roles but rather redefine and enhance them, contributing to reduced errors, shortened wait times, and overall improved quality of life for staff.

4. IT teams will need unified AI platforms

With the proliferation of AI agents, healthcare IT teams will face the challenge of managing a multitude of AI systems effectively. In 2025, the need for unified AI platforms will become evident; healthcare IT teams aren’t meant to be simply AI agent integrators.

Unified platforms will allow IT teams to oversee and coordinate multiple AI agents seamlessly. Such a centralized approach will ensure that various AI-driven tasks – from benefit verification to intake calls to patient follow-ups – are managed efficiently, maximizing the potential of AI without overburdening IT departments. This shift will require robust infrastructure and collaboration, but the payoff in terms of operational efficiency and patient/employee satisfaction will be substantial.

5. Multi-modal AI will begin to transform healthcare 

Multi-modal AI represents a significant advancement in AI, incorporating text, image, and voice capabilities to offer real-time, integrated interactions. During the webinar, I shared a demo in which ChatGPT showcased its ability to give step-by-step instructions for making coffee, illustrating its potential for more complex applications in a healthcare setting. Every foundational model developer is launching these and the quality will only improve over the coming quarters. 

In 2025, we’ll start to see the ways in which multi-modal AI can help solve major issues in healthcare, even for those that require in-person interactions. The ability to guide a patient through different steps in their care journey while on a video call with an AI agent will unlock time from seemingly routine clinical tasks. Using in-home injections as an example, multi-modal AI could assist patients in administering such a treatment themselves without a nurse present – something that could help expand access in remote areas or ensure patients could receive infusions even where workforce shortages are a challenge.

6. Web browsing agents will force a transformation of how pages are designed

Hundreds of millions of agents will be surfing the web as personal and work assistants in the coming year, forcing a shift in how we interact with the web. Data previously locked behind portals will be transformed into APIs, which means data access will become instantaneous. While these changes may not all occur in 2025, at a minimum, the groundwork will be laid.

Combining these changes with clinical intelligence will make processes like prior authorization much less burdensome than it is today. Instead of lengthy back-and-forths between healthcare providers and payors, these agents could instantaneously gather and validate necessary information, reducing administrative overhead and improving patient care delivery.

7. The business model for AI agents will be driven by measurable outcomes

Seat-based licenses are already a thing of the past. Modern companies charge for usage, and grow (or shrink) with their customers. But as AI agents continue to evolve, we’re on the cusp of yet another transformation: In 2025, companies will increasingly adopt outcome-based pricing models, where customers pay for the results delivered rather than the resources consumed.

This, perhaps surprisingly, will align interests. Unlike usage-based models, which can inadvertently reward inefficiency, outcome-driven models will naturally incentivize more efficient and effective AI solutions. With a shared focus on measurable success, healthcare organizations can more confidently invest in AI agents, knowing that their costs directly correlate with real-world benefits. 

AI adoption in 2025 and beyond

Also as part of the webinar, we dove into some real-world examples of how Infinitus customers are currently using AI – and plan to as we charge toward 2025. Healthcare AI in 2025: Trends, adoption, and opportunities is on-demand now if you’d like to learn more. Or, you can reach out directly to our team to kick off your AI strategy now.