How AI agents & generative AI redefine operational excellence

Artificial intelligence (AI) agents and generative AI are two of the biggest developments in the evolving AI landscape, and their ability to transform operational excellence (OPEX) is vast. AI agents are autonomous intelligent systems that perform specific tasks without human intervention, while generative AI has the power to create text, images, videos or other data using generative models.

By applying AI agents and generative AI to OPEX strategies, modern businesses can unlock significant advantages that were not accessible prior to recent technological advancements, says Nathaniel Palmer, CEO of Infocap AI Corp and the author of Gigatrends (2024), which recently reached #1 on Amazon’s “Hot New Releases” list for books on AI and machine learning.

An influential business process management (BPM) thought leader, Palmer has co-authored over a dozen books on BPM and process improvement, as well as being the first individual named as a “Laureate in Workflow.” Over his career, he has led the design and execution for some of the industry’s largest and most complex projects involving investments exceeding US$200 million and overseeing more than $2.5 billion in research and development around automation and AI.

PEX Network sat down with Palmer to learn more about the seismic changes AI agents and generative AI are introducing, and what it means for the future.

PEX Network: How will AI Agents and Gen AI redefine OPEX?

Nathaniel Palmer: This is happening already, but the opportunity we face as the process excellence community is shaping which direction this redefinition goes. AI is already ubiquitous, yet the disconnect between expectations and reality remains significant. That I am sure we can all agree on. What we are seeing now, as recent as weeks ago, is an evolutionary leap of generative AI from simply providing answers or generated content to the ability to act – moving beyond information to action. This is what is meant by agentic AI or AI agents, able to function as a digital worker capable of executing tasks, and even collaborating with human co-workers.

For AI to be trusted and enabled to act, this cannot take place inside of an algorithmic black box – it must be a “glass box.” Actionable AI needs to explain its reasoning clearly for every action it takes, rather than operating within a black box as it typically does today. By that I mean an unambiguously data-driven approach where decisions are based on quantifiable metrics rather than mathematical models and inferences – where every automated outcome is observable, explainable and auditable.

This is the key, which is overlooked in the growing hype and enthusiasm over generative AI and is a disaster in the making. Every future-focused leader getting excited about AI needs to hear this: If you want to leverage AI, begin with the framework of OPEX.

PEX Network: That’s a very interesting perspective, but isn’t it the same thing as ethical AI?

NP: Not exactly, but you raise a great point. Although there isn’t a single accepted definition, ethical AI largely relates to the use of data in training models and whether it reflects bias or has been expropriated without its owners’ permission. Here we are talking about ethical automation, or what we have specifically deemed human-centric automation.

Human-centric automation delivers capabilities which are human-like – where generative agents perform the same work as performed by humans but with the efficiency, accuracy and predictability only possible through automation. The potential for efficiency gains here are, of course, enormous. In our own work, we have seen gains as high as 500 percent in otherwise complex work previously deemed impossible to automate, with measured quality greater than 99 percent. 

AI-powered automation must transform processes to unlock potential in the human workforce, not by simply mimicking but extending human capabilities. The issue with human workers is that they are human – unpredictable, making decisions based on their own subjective interpretation of rules and policy, perhaps influenced by what they believe to be the right thing to do or simply most convenient for them. Yet as it turns out, what it takes to enable successful automation also makes the work done by humans easier and better (more consistent, predictable less reliant upon subjective interpretation of policies and rules). This simultaneously expands the aperture for what is automatable, where digital workers and human workers use the same systems, follow the same rules and are equally observable and accountable. That is human-centric automation.

PEX Network: Do you have examples of companies that successfully integrated generative AI into their practices?

NP: Absolutely. Two examples are DuPont’s leverage of AI for predictive maintenance and scheduling, and the Kansas City VA Medical Center’s use of AI to predict critical health events, enhancing patient care and operational efficiency.

Generative AI can be leveraged to better articulate problems and set project goals with greater precision by analyzing large amounts of data, such as customer feedback and operational metrics. DuPont applies it today to optimize production scheduling and sales pricing, as well as using agents to automate accurate and consistent data collection.

Kansas City VA Medical Center uses AI to predict patient risks and optimize care pathways, significantly improving accuracy in identifying when patients need critical interventions and reducing mortality rates as a result. We have worked with other customers in a similar manner, applying AI to solve otherwise impossible scheduling challenges involving thousands of service-level agreements (SLA) rules against hundreds of thousands of real-time data points, as well as to perform complex adjudication work with myriad rules and regulatory requirements upheld.

PEX Network: So generative AI is not just effecting OPEX, but OPEX is driving AI adoption?

NP: Exactly! It is time to look beyond the overplayed story that “AI changes everything.” Of course it does, but amid the rapid growth of AI and related technologies, the gap between ambitions and the ability to execute further widens. Transformation leaders need a partner who can translate their vision into actionable advantages – and that is the role of the process excellence community!