Augmenting, Rather than Replacing, the Knowledge, Skills, and Abilities of Experienced Staff with Generative AI, is the Most Promising Use Case in Healthcare
The following is a guest article by Mark Bates, Vice President of Product Management at Intelliworx
There’s a craft bar in Key West that harbors a secluded culture of brainy board games. These games often have esoteric rules that challenge one’s intellect. There are too many to remember, and yet mastering the details is essential to success.
The author of this article found himself playing a game against a man, from Generation X, who had a stunning memory. His ability to recall and apply arcane rules, often after hearing them only once, enabled him to put together game-winning strategies repeatedly.
By day, he was a construction worker. His specialty was applying Stucco, a fine plaster that’s used for finishing wall surfaces and molding architectural designs. Although he had no college education, his mind was so sharp that our company thought, with a little coaching, he might do well to fill a junior role as a quality assurance (QA) analyst open on the team.
He expressed interest when we first proposed the idea and hired him as a paid intern. He knew nothing about software development or quality assurance in the beginning. However, we have decades of experience on the team, and someone spent time mentoring him every day.
Access to SMEs Mirrors Access to Generative AI
A key to his role was understanding how to interview people for research and development (R&D). That is, gleaning insights from current and future end users of potential new software features our company is developing.
These aren’t typical interviews. The task is to draw forth business knowledge from interviewees, particularly around business processes. He needed to drill into the details and document what he finds so that a developer can turn the ideas into code.
What he lacked in expertise, he made up for with an insatiable curiosity, desire to learn, and a phenomenal intellect. When he had questions, a teammate served as his subject matter expert (SME). He could query us, and we’d give him the answer, while also pointing out lessons and pitfalls learned along the way.
The access he had to a readily available SME accelerated his learning and was an obvious boost to his productivity. Further, he’d often take what he heard in his interviews and from his mentors and combine those ideas in novel and interesting ways.
Initially, he needed a lot more guidance, but over time, as he gained experience, his contributions became more actionable. Over the course of three years, he has since become an exceptional quality assurance professional with a full-time job on our team.
It occurred to us, in a sense, that this is very similar to what businesses are striving to get out of generative artificial intelligence (AI).
The Three Ways Healthcare is Approaching AI Adoption
Like many companies in the tech space, we’re researching and experimenting with various generative AI use cases. As we continue to dig into the subject matter, we’ve observed there are three broad approaches healthcare companies are taking to generative AI adoption: wildcard, passive, and augmentative approaches.
1. Wildcard
The wildcard approach is a method of simply turning the AI loose on your data. The hope is that it will identify unique insights that an experienced employee can turn into productive ideas. It could produce something useful, or it could be a waste of time, budget, and computing power. The outcome is truly a wildcard.
2. Passive
The second approach is passive implementation. This is where an organization rolls out generative AI tools to all employees and leaves it up to them to find productive use cases. A key challenge is that generative AI presents answers with authority, even when it’s wrong. An experienced employee looks at the output and thinks ‘Generative AI doesn’t work very well.’ Conversely, the output might look fantastic to an inexperienced employee but is rife with redundancies or errors, which may reflect poorly on the business.
3. Augmentative
The final approach is putting generative AI in the hands of seasoned experts. It helps them gather and organize ideas and data, but their expertise is still required to determine what is meaningful, interesting, or otherwise actionable.
It’s this final approach that we are finding is most valuable. It complements the experience of seasoned staff, but it’s also reliant on their experience to add real value. This third use case is where we believe healthcare will get the most value out of AI.
Generative AI for Augmenting Healthcare Recruiting
We’ve been testing this approach in a variety of ways. One of the most urgent ways is to augment the productivity of human resource professionals and healthcare recruiters, amid a prolonged shortage of providers.
Take writing a job description for filling a physician vacancy, for example. It’s a time-consuming task one might be tempted to rush, but it’s also important. This is because a job description will make the first impression on a candidate. Research shows that about four in 10 providers will decline an offer if a prospective employer’s recruiting process seems disorganized.
We’ve tested generative AI for this specific use case across each of the three approaches. Generative AI on its own misses the mark. Further, we’ve observed that less experienced staff rely more heavily on the output. Neither approach will bring prospective providers to an interview.
By contrast, in the hands of an experienced recruiter, the job descriptions were unequivocally more enticing. While the recruiter nailed down the technical details of the role, the AI also suggested new details the recruiter hadn’t thought of and was keen to include.
The AI didn’t replace the recruiter – it augmented their work. Importantly, what makes this use case productive is that the recruiter had the experience to recognize the value of the AI-generated suggestion.
Leaders should keep in mind that this is just one example. There is a range of possible use cases where generative AI can augment your existing staff and increase productivity and efficiency without sacrificing quality.
Augmenting Experienced Healthcare Professionals
News, by definition, is something that defies expectations. That’s why it’s the exceptional case studies – those that make generative AI seem like magic – that make the most headlines.
Yet when technologists experiment on their own, the results are lackluster, which contributes to a growing disillusionment with the return on investment of AI projects. The internet is filled with memes and screenshots of where generative AI has steered them wrong.
This is why we argue, philosophically, that the best use case is augmenting existing skills, knowledge, and experience, especially in a nuanced industry like healthcare. While generative AI can process and organize vast amounts of information – beyond a human capacity – it still takes a layer of institutional knowledge to identify the kernel of insight worth pursuing.
Augmentation is the use case software providers should be pursuing in developing solutions for healthcare organizations. Similarly, it’s also the best use case for healthcare organizations to evaluate for procuring and deploying software.
No comments:
Post a Comment