Artificial intelligence (AI) has exploded onto the scene in most industries. And the healthcare field is no exception. Faced with growing workloads, shrinking margins and the ever-present “do more with less” mindset, many in healthcare are looking to generative AI as the next revolution in healthcare delivery, management and marketing.
We’re bullish on the concept of generative AI. But it’s important to not consider the technology as some “holy grail” or “silver bullet.” In a prescient presentation given at IBM in 1979, there was a slide that read: “A computer can never be held accountable. Therefore a computer must never make a management decision.” Any tool, no matter how powerful, is only as good, helpful and considerate as the person who wields it. And there are many considerations for AI in healthcare organizations — everything from HIPAA, PHI and PII to being mindful of the consumer experience and even finding ways to lower the risk of burnout among your organization’s providers.
We’ve broken down six of our recommendations for best practices your team should keep in mind as you determine how to incorporate AI into your organization’s operations.
Do establish an organizational AI policy
No matter if you intend it or not — no matter whether you even know it’s happening — there are almost certainly people within your healthcare organization right now who are finding ways to use AI in their tasks. From creating reports and managing data to writing messages, AI is a powerful method of taking large, unruly inputs and generating functional outputs quickly. But therein lies the danger as well as the strength.
Working with untested, unsupported tools is dangerous from the security and system liability standpoints. Without careful integration, an unmanaged AI tool could lead to serious violations regarding patient data, your organization’s internal data and other vital information. Employees who “go rogue” and start using unvetted, unmanaged AI tools can open your organization up to serious liability.
The best time for your organization to have set an AI policy was yesterday. The second-best time for it to do so is right now. This kind of policy gives you tools to safely introduce generative AI systems into your workflows. And the policy will provide guidelines to your team members about when to use — and not to use — AI in their duties and processes.
Bringing in a qualified healthcare data management team with experience in AI implementation is key. These experts will work with you to integrate AI into your processes safely and efficiently. This can help your team members can work not only smarter and faster but also safer. And an AI data management team can help ensure your implementation aligns with strict requirements from the Department of Health and Human Services (HHS) and other regulatory bodies.
Don’t just use any AI system
Not all AI systems are created equal. And by the very nature of AI and machine learning (ML), none of them evolve equally either. So AI is very much a tool and solution that requires a customized, thoughtful implementation in order to be successful.
If your team members just select a generative AI system and expect an easy plug-and-play experience, that’s a recipe for varying levels of disaster — from bad user experiences up to and including potential liability due to regulatory noncompliance. Even firms who know AI and know it well don’t always know what it takes to implement AI solutions in the healthcare industry. So it pays to work with experts in both AI and healthcare.
When you establish an AI policy and then work with your generative AI partner to source and implement the system carefully, you minimize your organization’s risk for these negative — and potentially costly —experiences.
Do get input on where generative AI can lighten the load
It’s no secret that healthcare organizations took a substantial hit during the COVID-19 pandemic. And they’re still recovering, having to do more work with fewer people and less resources. We’re continuing to see recruitment and retention as major issues for our clients. Organizations are having trouble aligning workloads to their headcounts.
This is one of the key ways generative AI can help you. From lessening the burden on your call centers to helping providers shoulder the burden of charting and responding to patients notes in Epic and other EHRs, generative AI works best when it’s used to speed and streamline large-scale, repetitive tasks. But it’s not a one-size-fits-all solution.
We recommend putting together a focus group of representative users — not just “power users” or people in your organization who are already using generative AI. Seek insights from department heads, providers, administration and leadership about where AI solutions could make their everyday tasks easier.
We also recommend establishing an AI committee to work cross-functionally with your departments. Ideally, this committee would include representatives from your IT and legal departments. This will help you accomplish several key tasks:
- Consolidate requests for AI implementation to avoid duplicating work
- Determine the feasibility of various requests to use AI tools and whether they comply with HIPAA and other regulations
- Help lead the charge for departments and leaders who may be resistant to AI usage
- Find new ways to incorporate AI that can benefit multiple departments (e.g., if it worked over here, could it work similarly over there?)
Don’t think of AI as the one-stop shop for your content strategy and content marketing
We’ve heard this from clients who are daunted by mountains of content work before them: “We’ll just farm that out to an AI.” It seems like the easy answer for content concerns. Generative AI can write (sort of), and people barely read healthcare content anyway, right?
It’s true that most healthcare consumers aren’t reading every letter, word and comma your content team writes. But they do skim and scan for what they’re interested in. And what they’re interested in is exactly the material generative AI is terrible at creating for you. Say it with us: “Generative” does not equal “creative.”
Your healthcare content depends on showing consumers what makes your organization special — the differentiators that set you apart from your competitors. And generative AI can’t account for those. It was trained in the Wild West of the internet — from Reddit and X (the site formerly known as Twitter) to blogs, personal websites and more. All that data the large-language models (LLMs) that power generative AI programs have crawled means they can write based on patterns of what words are likely to come next in a given phrase. But the AI doesn’t know and can’t comprehend your value proposition, because that’s unique to your organization. And it can’t speak to your healthcare personas, their unique journeys or their distinct concerns and objections.
And, in a key consideration for healthcare especially, AI cannot write with empathy for the healthcare consumer’s journey. It can fake empathy — and it’s better at faking it than ever before. But, at the end of the day, generative AI programs are still just programs. They can’t understand emotion and, therefore, can’t write with an understanding of what your consumer is feeling. Good healthcare content should come across as caring, supportive, knowledgeable and friendly — traits your organization should embody. But AI-generated content can come across as cold and distant. And that means your potential patients or members can erroneously think your organization is cold and distant too.
Generative AI can help if you need to explain a pancreaticoduodenectomy or to describe the general process of a coronary revascularization. What it can’t do is explain why consumers should trust your team to help them if they need these procedures. It’s important to use the tool for what it’s good at, not try to force it to do something it can’t — or shouldn’t — do.
Do use AI to improve your search and chatbot interactions
Site search used to be almost an afterthought for most digital implementations. The idea was that people got where they needed to be from Google, Bing or other search engines and then would find their own ways from there. But as more consumers expect a robust search tool as table stakes (blame the Amazon effect!), search has grown ever more important in the digital experience.
Generative AI is delivering strong semantic and conversational search experiences that you can integrate into your site search. This is leading to more relevant search experiences and better user engagement for clients who make these investments in their search solutions.
Generative AI is also elevating chatbots — once limited to a handful of scripted interactions — to the level of true conversational assistants. AI-powered chatbots, when created and implemented thoughtfully as part of a consumer-centric experience, can adapt to unpredictable user queries and needs on the fly.
We strongly recommend including user testing when implementing or upgrading your AI-powered search or chatbot solutions. Get feedback from your actual users (or users in your desired demographics) to find what’s working, what’s not working and how users would respond to potential improvements.
Do choose the right generative AI partner for your organization
The potential of generative AI is real. But so are its perils. You need the right partner to help you both formulate your organization’s strategy for how to use AI and then implement that strategy.
Perficient’s experts have helped lead the charge in the AI transformation, particularly within the healthcare space. We’ve worked with clients across the country and around the globe on thoughtful, strategic AI implementations. Generative AI is playing an important role in the strategies we’re offering for clients in data service, modernization and business transformation — key pillars of our practice and for your healthcare organization.
What AI looks like now in the realm of healthcare won’t be what it looks like in a year, five years or 10 years. But AI is here to stay. We can help you work through the considerations, goals, KPIs and implementation process. And we’ll be here to help you create a strategy that will grow and mature with your organization over the long haul.
We’d love to talk with you about your AI questions, challenges, concerns and goals. Contact us to learn more about our healthcare-focused generative AI solutions.
