Jun 23, 2025

AI innovation is accelerating at a pace unlike anything we have seen before. Ethan Mollick, a well-known Wharton professor and thought leader on AI, put it best when he said that today's AI tools are the worst they will ever be . Meaning, while we've already seen AI tools advance significantly, AI is only getting better, faster and smarter, and we need to get ready for the AI capabilities of the future. In today's AI era, organizations are participating in a massive delegation exercise, trying to figure out which tasks can be fully delegated to AI and which tasks require human oversight. However, leaders and AI decision-makers must keep in mind what AI's future potential could look like when strategizing today. New research conducted by Bessemer Venture Partners, Amazon Web Services and Bain & Company shows that over half of healthcare organizations are seeing material ROI in 12 months of AI implementation. For even more healthcare organizations to maximize ROI, leaders must recognize that they should not simply replicate their existing workflows with AI. AI implementation must instead be a redefining process where leaders take a daring and holistic approach to match the pace of AI innovation. While some solutions may not be possible today, they may be in the future. Through this exercise, teams will identify where human intervention is necessary and establish appropriate guardrails and AI governance mechanisms. Organizations that take this approach will stay ahead of the AI innovation curve, enabling teams to build more effective workflows that allow employees to focus on high-value work that drives business growth. Enable employees to activate the AI strategy. AI is most powerful when organizations use it to redefine operations. Big organizational shifts are driven by small, incremental changes. Leaders should enable teams to think creatively about AI implementation and encourage employees to envision what a completely automated workflow could look like. If teams do not imagine what is possible without restrictions, they are unintentionally putting up barriers and stifling innovation. By coaching employees to imagine a blank canvas and design a fully automated workflow, organizations can uncover new opportunities for AI. For example, at my organization, Arcadia, an employee noticed colleagues asking similar questions across internal Slack channels, recognizing an opportunity to use AI to automate a process where humans were not needed. He created an AI tool that connects to multiple internal sources and provides automated answers to employees' questions in seconds, reducing the amount of time employees spend finding answers for each other and enabling them to remain focused on higher-value tasks. These types of low-risk, high-reward tools deliver an efficiency improvement of approximately nine orders of magnitude and underscore how AI-powered innovation is possible when leaders encourage teams to think outside the box. Integrate AI systems to maximize effectiveness. Once teams understand how fully automated workflows might operate, they can then start building the foundational requirements for optimization. A great example of an organization that is leaning into AI implementation is Kaiser Permanente, which recently published a case study on their AI-powered scribe rollout—including lessons learned. Among those takeaways, Kaiser surprisingly changed vendors between the initial pilot and final rollout of their AI scribe because they wanted the tool to be better integrated into their electronic health record system. As I have written about before, when AI is integrated into existing workflows, the outcomes are always better. At Kaiser, providers liked the initial AI scribe, but agile leaders pivoted quickly because they wanted the tool to work even better. They saw the full potential of this technology and decided to drive more meaningful impact for their business. To build the foundation for automated workflows, connecting tools that use various types of AI is key. In the Kaiser example, the AI scribe tool they ultimately chose uses agentic AI, which autonomously makes decisions and solves multi-step problems without human assistance, to transcribe a patient visit. Generative AI, which requires more human intervention as it solves less complex problems while relying on prompts to produce specific outputs, is layered in to produce a summary from the transcript. By stitching together tools that leverage different types of AI, teams are able to lay the foundation for the fully automated workflows of the future, while also helping eliminate administrative, time-consuming tasks and enabling employees to focus on the highest value work. Understand where human intervention is critical. For now, leveraging a more holistic AI implementation strategy does not mean that humans are fully out of the loop. I intentionally write “for now” because it's difficult to predict how dramatically AI could transform our world. But today, this means that organizations have better insight into instances where a human touchpoint may need to be more heavy-handed. Many still look at integrating AI from an extreme perspective—either humans do everything or machines do everything. But I view the degree of human intervention as a spectrum. In healthcare, primary care delivery is an area where doctors should not be out of the loop (for now), but AI can help automate administrative tasks, like developing patient summaries. However, with many large pharmacy chains struggling financially , an example of an area primed for automation is pharmacy workflows , where automated prescription drug dispensing can help reduce the potential risk of error. When organizations have a deeper understanding of where human intervention is needed within workflows, they can establish more effective AI systems. Leaders and AI decision-makers must take teams out of their comfort zone and encourage them to think differently about AI implementation to match the unstoppable pace of AI innovation. Real transformation will take place when organizations implement AI tools holistically and stitch them into a cohesive system. This way of thinking will empower companies to make more strategic AI investments and harness the power of AI, driving more efficiency and growth for their business. Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?