I’ve been asked many times lately about how to identify improvement opportunities for AI and how to measure the success of an AI initiative. I usually respond something like “How did you do it before and why should this be different?” Artificial Intelligence (AI) can feel like something that arrived overnight. It didn't. AI has been evolving for decades. Businesses have used machine learning, predictive models, optimization algorithms, and other forms of intelligent automation for years. What has changed dramatically is what these technologies can do, how accessible they have become, and how quickly they can be put to work. That makes this an excellent time to explore AI, but we shouldn't forget what we already know about improving a business.
AI changed nothing about how we should look for opportunities.
Start with outcomes. What do we need to accomplish that we struggle with today? Then understand the process, break it into its component activities, and ask which parts require a person and which could be performed or assisted by a machine. Where are we spending too much time? Where are we experiencing delays, errors, downtime, or quality issues? Where are we experiencing the most employee turnover? Where are people doing work that adds little value? Don't start with, "Where can we use AI?" Start with the outcome you need and the problem standing in the way.
AI changed nothing about ROI.
When someone asks "How do I calculate the ROI of an AI project?", I respond “The same way you calculate the return on any other improvement.” Establish the baseline. How long did the process take before? What did it cost? How many errors occurred? What was the throughput, downtime, scrap, conversion rate, or customer response time? Make the improvement and measure again. The technology may be newish, but the math isn't. If ROI isn’t important to you, that’s fine. Measure what matters to you, but still start with a baseline for your benchmarking.
AI changed nothing about who is responsible.
Despite headlines suggesting that AI is “taking jobs,” a better way to think about what is happening is that we are delegating more tasks to machines than before. We have been doing that for generations. A calculator took over arithmetic, industrial robots took over repetitive motions, software took over countless administrative tasks, and now AI can take on activities involving language, images, analysis, and decision support. But delegation does not eliminate accountability. If I delegate a task to a coworker, I am still responsible for ensuring the work is done correctly. The same is true when I delegate a task to AI. People must still define the outcome, provide the right context, establish appropriate guardrails, validate the result, and remain accountable for what happens next. AI may change who (or what) performs the task, but it doesn't change who owns the outcome.
AI changed almost everything about what can be automated.
Traditional automation worked best when inputs were structured and rules could be clearly defined. Generative AI dramatically expands that boundary. Technology can now interpret an email, summarize a document, extract information from an invoice, analyze an image, search thousands of pages of institutional knowledge, draft a response, or turn a plain language request into software code. Activities that once required expensive custom development or simply couldn't be automated reliably are increasingly within reach. When people are considering use cases appropriate for AI, I tell them to avoid the temptation to discern whether AI can do it or not. Rather, imagine that anything is possible to delegate as long as you can describe the process as if you are delegating it to another person, and have a way to know if the work was completed successfully or not. If you identify a task that could not be delegated to AI, that’s great. In that case, you are learning what is and what is not a good fit.
AI changed how we talk about innovation.
For years, conversations about innovation often stayed within technology departments, engineering teams, or Continuous Improvement teams. AI changed that almost overnight. Today, executives are asking about copilots, supervisors are experimenting with AI assistants, employees are finding ways to eliminate administrative work, and organizations are asking what autonomous agents might eventually accomplish. People who never considered themselves technologists are suddenly bringing technology ideas to the table. That may be one of AI's most significant contributions: it gave nearly everyone permission to participate in the innovation conversation.
AI changed how we should view some of our oldest challenges.
A workforce shortage isn't only a recruiting problem anymore. It is also an automation opportunity. High turnover in a repetitive position should make us ask whether every task in that job still needs to be performed by a person. Unsafe work should cause us to consider whether machines, robotics, computer vision, or remote monitoring can reduce human exposure. A technician making rounds to manually check equipment might instead be supported by sensors, Industrial Internet of Things (IIoT) connectivity, condition monitoring, and predictive analytics. A repetitive material handling task might be a candidate for robotics. Quality inspections might combine machine vision with AI to help people identify defects sooner.
That's where this conversation becomes much bigger than generative AI.
The real opportunity is the convergence of AI with industrial automation, robotics, sensors, IIoT, machine vision, connected equipment, data analytics, and the people who understand your processes. Individually, many of these technologies aren't new. What's new is how accessible they are becoming and how powerfully they can work together.
AI changed nothing about the need to understand your processes, define outcomes, measure results, and solve the right problems. But it changed almost everything about what's possible once you do.
Let us help you on your journey.
If you would like to learn more about how AI and other emerging technologies are changing the workplace, join us September 22 at 3:30 at the WKU Innovation Campus for the launch of the Smart Technologies Peer Group. See https://www.regionaltechcouncil.org for more information and to RSVP for this free event.