Over the next three to five years, AI adoption is likely to remain centered around larger and academic health systems that tend to have the size and resources to leverage emerging technology. Nonetheless, within these larger organizations, adoption is still often experimental, and it will take some time before AI is deployed at scale. Meanwhile, small hospitals and health systems are less likely to be pioneers in this space, and are usually adopting a "wait and see" approach.
The path to a successful AI transformation
So what does a successful transition to AI look like? A McKinsey Global Institute discussion paper recently looked into this, outlining a framework for how organizations can navigate the AI adoption process:
- Start by evaluating use cases and creating the business case
- Establish a data ecosystem by breaking down data silos and sourcing/refining high-quality data
- Identify fit-for-purpose AI tools, partnering or acquiring capabilities as needed
- Integrate AI into workflow processes to optimize the human-machine interface
- Build and foster an open, AI-friendly culture
McKinsey went on to define traits of organizations (across multiple industries) that have had some early success in this AI transformation.
For example, McKinsey found that on average, larger organizations that had a history of digital adoption are more likely to adopt AI. For these digitally mature organizations, AI is simply the next wave of digitization, and they had an advantage in terms of scalability. By contrast, organizations that have been slow to adopt digital technologies tend to trail the pack. In addition, early adopters typically do not fixate on one type of AI technology. Rather, they adopt multiple AI tools that address a number of different use cases and then scale up the pilots that show the most promise. Furthermore, early adopters tend to be motivated by the upside growth potential of AI instead of focusing solely on cutting business costs. In other words, AI is not only about process automation for these organizations; AI is a key driver of major product and service innovation.Characteristics of Early AI Adopters
Here are some initial steps you can take to prepare for AI's impact in health care:
- Make plans now: Evaluate your IT infrastructure sooner rather than later. There are no shortcuts to AI, and early adopters are gaining competitive advantages. Consider leveraging partnerships with universities, research centers, corporations, and other health systems where appropriate.
- Smaller organizations have options too: Apart from buying vendor products that may already have AI "under the hood," there are growing opportunities for small-scale organizations to leverage open-source AI tools (e.g., Health Catalyst's healthcare.ai project) and knowledge repositories to test out pilots and gain some fundamental insight into how AI can be applied for innovation and competitive advantage. For example, Zebra Medical Vision, an Israeli-based imaging analytics company recently announced it will offer access to its image-interpretation algorithms for just $1 per imaging scan.
- Optimize internal and external expertise: Have access to the right specialists who can bridge IT and business needs. Deploy intelligent systems with the active input of frontline staff in the organization to incorporate their knowledge and experience and reduce resistance to new IT systems. Executives and other non-IT leaders probably don't have to know the intricate details of how AI works, but should take some time to learn the basics, and have an intuitive understanding of AI.
- Learn from others: Pay attention to how AI is affecting certain job functions, staffing, workflows, or cost structures, not just in other health care organizations, but in other industries. Even if you're not going to be an AI pioneer, it's important to be aware of AI disruption and be ready to adapt your business model.
Learn how machines analyze emotions to improve patient care.