Maximizing AI ROI In healthcare by establishing an automation-first mindset [Q&A]
AI use cases in healthcare continue to expand, and organizations are identifying opportunities to leverage automation technologies to improve existing workflows.
However, after years of implementing new tooling and expanding tech stacks for better efficiency and backend processes, most organizations are left with disparate systems and datasets that make AI initiatives difficult to put into practice. IT teams are now forced to work backwards, embarking on time and resource consuming efforts to determine how automation can work within the parameters of these fragmented processes.
We spoke to Rob Duffy, chief technology officer at HealthEdge, about how to establish and embrace an automation-first mindset to become more agile, optimize operations and uncover opportunities for innovation.
BN: What’s the first step in establishing an automation-first mindset in a healthcare organization that spans across clinical, technical and business teams?
RD: The first step in any major organizational change initiative, regardless of size and scope, is to gather support from the leadership team. The same goes for becoming an automation-first healthcare organization and embracing AI. The healthcare industry has its own unique set of financial and operational complexities but foundationally, this is a change management activity. Senior leaders may require some convincing on how this will address their biggest concerns, so it’s best to articulate how this will reduce costs and clinician burnout, improve patient outcomes and member satisfaction, and a realistic timeline for a phased roll out or implementation. It’s also important to note which departments will be involved and how their ongoing workflows or processes will be impacted during the transition. Cross-functional and multidisciplinary participation are critical to organizational-wide adoption
Once leadership has committed to an automation-first shift, the next move is to determine which processes should be automated first. This means consulting with key stakeholders to identify teams most open to adjusting how they currently operate, workflows mired by cumbersome manual tasks and what changes will have the greatest benefit on cost and care. From there, set attainable goals with low-risk, high-reward activities and implement automation to reduce human intervention and still accomplish them. Starting small and successfully executing will make it easier to establish an automation-first mindset across the clinical, business and technical teams.
BN: What role do factors such as regulatory compliance, data security (particularly surrounding PHI) and potential impact on patient care play when deciding which processes to automate in a healthcare setting?
RD: The criteria for selecting initial automation projects must be carefully considered given the highly regulated nature of healthcare and critical importance of data privacy and patient safety. It should account for the use-case value, operational readiness and risk of integrating technologies. Healthcare isn’t an industry where we can simply add AI to products or platforms, automate every backend process or use an LLM to answer all our questions. The risks are far too great, and concerns surrounding data security and integrity are warranted because of the cascading effect they have on patient outcomes and member experiences.
CTOs and CIOs at healthcare payer and provider organizations are under immense pressure to improve efficiency and optimize existing resources. Instead of looking at AI and automation as an afterthought or bolt on solution to ongoing challenges, they should be foundational elements of a broader technology roadmap. This helps determine where they should begin their automation journey and inform how it will evolve. The roadmap starts with establishing governance policies and guardrails that ensure safety, transparency and compliance with state and federal regulations. The appropriate governance models allow technology leaders to analyze potential new workflows and solutions to which automation and AI can be applied, not simply incorporating the technologies into existing processes.
BN: Beyond generic benefits, what are some healthcare-specific use cases where automation and AI can demonstrably improve efficiency, reduce errors and enhance the patient or provider experience?
RD: Healthcare organizations are already automating administrative tasks and applying AI for internal productivity in areas that include scheduling, clinical documentation and patient engagement. The data being collected is typically well-defined and allows many payers and providers to use robotic process automation (RPA) tools. Digital patient intake forms, for example, are standardized and allow health systems to automatically create electronic health records that can be updated with additional data shared through chats, emails and other interactions. This allows clinicians to spend more time with their patients, eliminates the need for manual data aggregation and gives payers, providers and patients a holistic view of their healthcare journey.
As healthcare organizations look to reduce manual processes with AI and automation, the next phase of this journey is implementing systems that can complete non-clinical, abstract tasks and reduce human intervention when the data isn’t as well defined. For example, if a provider dataset is structured differently than a payer’s, someone has to go in and make manual adjustments to advance the process. There is tremendous value in creating agentic-led systems to execute ordinary workflows. These agents don’t require access to regulated data and won’t expose PHI to any outside entity or LLM. They complete the same tasks a human typically does, such as retrieving a file, querying a system and understanding how to navigate unstructured data.
BN: Given the rapid advancements in AI and automation technologies, and the specific needs of the healthcare industry, what can organizations do to ensure they don’t fall behind?
RD: The most successful digital transformation initiatives I’ve seen weren’t about the technology – they were about the people leading the change and adapting to the challenges faced on the journey. Healthcare organizations must prioritize AI and automation within their change management strategy. It can’t be the last initiative on an ever-expanding list used to accelerate the successful implementation of everything ahead of it. It needs to be a fundamental component and become muscle memory across all departments and functions.
Creating that muscle memory starts with AI literacy and encouraging internal stakeholders to adopt AI and automation. Once they have that in place, they’ll begin thinking about how they can solve existing problems and roadmap future processes using AI. It encourages an environment of continuous learning and innovation, and forces stakeholders to ask themselves if they would approach a problem differently or find an alternative solution -- faster -- if leveraging AI was a primary consideration.
Image Credit: BiancoBlue/Dreamstime.com