Harnessing AI Abundance: Strategies For Effective Implementation And Impact

Michael Meucci is President and CEO of Arcadia, a healthcare data platform that helps payers and providers put their data to work.

AI is top of mind for executives across every industry, with a 2023 Bloomberg report predicting that generative AI could become a $1.3 trillion market by 2032. As AI tools continue to pop up and evolve, we must consider the impact of the "abundance" of AI. On one hand, the abundance of AI is a net positive, particularly with many smart people innovating with AI to make business more efficient. However, the abundance of options and the volume of output from AI models can be overwhelming.

As Herbert A. Simon, a revered American scholar, once said: "A wealth of information creates a poverty of attention." When scaling AI tools, leaders need to do so with a focus on strategic implementation and managing abundance to create real change for their business and employees—who have been forced to drink from a firehose so far.

Defining AI Abundance

I define abundance in two different ways. First is the abundance of AI solutions. We've seen a surge of new AI tools brought to market over the last five years, as Stanford University's 2024 AI Index Report found that the number of AI patent grants worldwide increased by 62.7% from 2021 to 2022.

The number of partnerships between AI innovators and platform vendors has also increased, as they often view collaboration as an opportunity to ease the implementation burden caused by the abundance of AI. We have already seen this in my industry (healthcare).

For example, Epic plugged Abridge's AI capabilities into its EHR system, which allows providers to record and summarize patient conversations and integrate information into the patient chart. In this case, Abridge's AI tool is more effective when integrated into the existing workflow rather than operating as a standalone point solution, which drives adoption and benefits the end user.

Second, AI tools can produce an abundance of output. The prompt for AI tools must be specific; otherwise, the end user will end up with a mountain of information—which exacerbates inefficiencies. That said, even when using a targeted prompt, AI tools may still provide outputs that might not materially change how someone operates. For example, email made communication more efficient, but we still have to sift through our inboxes to find relevant messages. We've trained ourselves to ignore the noise, and we need to do the same thing with AI.

Integrate AI Into Existing Workflows

To better manage AI abundance and more effectively deploy AI, organizations must embed AI tools into existing workflows. In doing so, organizations can shift from leveraging passive AI (i.e., a notetaking or summarization tool that runs in the background) to active AI that users can truly interact with.

Ultimately, convenience drives adoption. For example, many people use ChatGPT to write or edit emails, but it is not integrated into Microsoft Outlook. Conversely, Microsoft Copilot is directly integrated into Microsoft applications. Workflow-integrated tools will come out on top; even if a more advanced tool exists that's a better fit, convenience will trump perfection to drive adoption in most cases.

When looking at healthcare, an industry where AI shows significant promise, if AI is integrated directly into the EHR, I believe it could become the co-pilot needed to empower providers so they can make appropriate clinical decisions and drive high-value, low-cost care for patients based on their unique physiology. We are already seeing Oracle put this into action, as it has previewed its next-generation EHR designed to leverage AI across the entire workflow.

How To Act On AI Abundance Now

Strategic And Selective AI Implementation

While every organization is looking to innovate using AI, moments will arise when AI is not the right solution. In these cases, it is best to empower your people to determine which technology is best suited for the use case. Sometimes, an existing solution is truly the right choice. In other words, there's a reason why we've used the wheel for over 6,000 years.

Choosing The Right AI Model

If an AI tool is the best solution, selecting the right AI model is critical. An incredible amount of AI tools are on the market today, but deciding which AI tools are the best fit and what is the best way to use them is not always clear—leading to ineffective use of AI.

Users must also know what blind spots these tools have. Large language models are growing rapidly, and an off-the-shelf AI model is all-purpose. However, some organizations need a specialized model for specific use cases. I see this when working with healthcare organizations that need AI models that are fine-tuned with greater clinical knowledge.

High-Quality Data Is Key

Our research shows that 61% of organizations refresh their data at least daily for business intelligence analytics, but when building and running AI models, that number drops to 32%. High-performing AI tools are built off high-quality data, and continuously training and updating AI models with fresh data helps produce the most current and accurate outputs.

Increasingly, organizations are looking to data and analytics platforms as well as data unification solutions to help improve their data quality for this purpose. In addition, organizations can train AI models to include local data to make outputs even more accurate. For example, healthcare organizations may opt to train AI models using local patient demographics or health patterns to make the output more relevant and ensure it reflects the real-world dynamics of the patient population they care for.

As leaders examine their organization's AI roadmap, they must have a user-first mindset and not get distracted by shiny objects. Integrating AI tools into existing workflows can help make AI more user-friendly, which drives attachment and stickiness.

Looking ahead, I'm encouraged by how the tide is turning in healthcare around AI. More and more people at every level of the healthcare ecosystem are realizing how impactful AI can be when it is integrated into existing workflows, and I hope to see these changes improve healthcare delivery.

Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?