If you are like me, you have questioned the continued reliance on paperwork and procedures involved in receiving care. Artificial Intelligence (AI) is often recognized as a promising avenue to address such inefficiencies by reducing patient wait times, enhancing the physician-patient experience, and streamlining workflow steps throughout the patient journey. However, these advancements have not yet been fully realized. This article presents several examples emphasizing the importance of involving patients in the development of AI solutions to support broader adoption and improved care delivery through AI-enabled technologies.
Generative AI and the Hype Cycle
Generative AI is beyond the hype cycle and is making its planning rounds within the board rooms and strategic planning decisions. The impact is varied across industries and a McKinsey and Company study indicated the economic potential of Generative AI, highlighting that more than 75% of the use cases falls under four main areas: Customer Operations, Marketing and Sales, Software Engineering, and R&D, to the tune of 2.6 USD to 4.4 trillion USD across 60+ use cases surveyed in this article. (Michael Chui et al, 2023, titled the economic potential of generative AI: The next productivity frontier)
source: McKinsey & Company, the economic potential of generative AI
AI Adoption in Healthcare
While the growth and prospect of GenAI is promising, adoption of such solutions in the healthcare industry is slow, primarily due to the lack of integration of such tools into clinical workflow. Recent work by Willemijn E. M. Berkhout, Msc et al. (2025) in JAMA Network Open found that only 25% of AI solutions progressed towards clinical implementation. The study found various factors such as lack of external validation, poor workflow integration, insufficient business cases as some of the reasons for the gap between the surge in AI research that does not always translate to successful clinical implementations. However, in my experience, I do believe that not having the patient’s use cases at the center of the AI journey slows adoption too. (Willemijn E. M. Berkhout, Msc et al, 2025, titled Operationalization of Artificial Intelligence Applications in the Intensive Care Unit)
According to the study, AI solutions implemented independently from the broader healthcare ecosystem often experience slower adoption. Gaining a thorough understanding of patient journeys during the development of AI solutions may contribute to improvements in current workflows.
Drawing on my experience with product lifecycles, I recommend that AI developers incorporate three key steps into their development process to ensure meaningful patient consultation when designing Generative AI solutions. There may be additional approaches, and I invite feedback regarding how your organization integrates patient perspectives during the development, deployment, or use of Generative AI in healthcare settings. Note: These recommendations may be applicable to various types of Generative AI solutions.
- AI-enabled software and services. Examples of such services include patient engagement & virtual care, patient monitoring, diagnostic imaging aids and clinical decision support tools.
- Co-pilot type software that helps in task automation. Examples of such services include clinical documentation and note-taking, coding and billing automation, and clinical workflow assistants.
- Advanced AI solutions that can be deployed as autonomous or semi-autonomous agents. Examples of such services include diagnostics and imaging agents, clinical monitoring and decision agents, robotics and autonomous procedure support, and patient-facing agents.
1. Consulting with patients as the anchor of “lived experiences”
A critical step in improving healthcare delivery involves the user experience related to how AI results are communicated to patients and caregivers. For instance, AI could assist cancer patients by sending follow-up reminders or generating reports designed to reduce anxiety. The development of such solutions may benefit from input provided by individuals who have experienced these conditions firsthand. The concept of incorporating the experiences of patients is becoming more prominent and should be considered in the design of GenAI tools.
2. Understanding the patient journey
A survey conducted by NEJM AI indicated that 91% of patients prefer to be informed when artificial intelligence is involved in making decisions regarding their healthcare. In contrast, 96% of respondents expressed concerns about data being used without patient consent. Looking ahead, it is anticipated that AI-enhanced reports and care decisions will complement human expertise. As requirements for responsibility, trustworthiness, and impartiality in AI persist, the integration of these technologies is expected to result in outcomes that are seamless from the patient’s perspective.
Although patients express a desire to be informed, an important consideration is whether simply stating “AI was used in your diagnosis” would enhance their understanding or facilitate the adoption of AI, particularly if patients are not engaged early in the process. Incorporating patients and considering their experiences in cases where AI-generated results are presented directly to them may help address obstacles to acceptance. (Dhruv Khullar, MD, MPP et al., 2022, “Perspectives of Patients About Artificial Intelligence in Health Care”)
3. Collaborate with Health Systems’ AI Governance Committees
In the United States, health systems are beginning to implement AI Governance Committees. According to a survey conducted by KLAS, only 16% of 35 healthcare facilities reported having a formal governance policy addressing AI use and data access; however, this proportion is increasing. The practice of convening comprehensive cross-functional committees to evaluate systems and processes is well-established in healthcare, as demonstrated during the adoption of Health IT and EHR initiatives. Recent regulations issued by the U.S. Department of Health and Human Services (HHS), which standardize data elements for health equity and transparency, represent significant progress toward building trust among patients and caregivers regarding the sources underlying AI-driven analyses. (Final Rule, Health Data, Technology, and Interoperability: Certification Program Updates, Algorithm Transparency, and Information Sharing by the Office of the National Coordinator for Health Information Technology (ONC), Department of Health and Human Services (HHS))
As hospitals establish AI Governance Committees, a critical consideration may involve including patients as part of the team to define policies for steering, governance, and ongoing monitoring of AI solutions. Utilizing Health Systems’ Governance Committee playbooks can support the Healthcare Generative AI ecosystem by ensuring adoption principles are considered during concept development and early testing stages. (Heather Landi, 2024, titled “Health systems struggle to put AI governance policies in place to keep up with tech innovation”)
Summary
Generative AI and artificial general intelligence (AGI) have significant potential in advancing healthcare. By integrating the expertise of clinicians and caregivers with the lived experiences of patients, AI-driven solutions can contribute to improved patient outcomes and help shape a more effective healthcare system for the future.
Thanks for reading. I welcome your thoughts and feedback.
Author
Venkat Ramamurthy is a distinguished executive with more than two decades of experience in progressive general management roles, driving growth through artificial intelligence and innovation within the HealthTech, MedTech, and Life Sciences sectors. In his current capacity as an advisor to several startups, Venkat provides strategic guidance to help refine product offerings and strengthen market presence.
LinkedIn profile page: https://www.linkedin.com/in/veram
