AI Panel (CONT’D FROM PAGE 1)
get started is to identify a tedious task that people dislike doing and try to automate it. Let’s dig deeper into that. Alvarez: Not too long ago, we implemented a docu- ment management system. It reviews invoices, indexes them, and routes them for approval. There is AI involved, but the main value came from streamlining the process. Zlatic: Change management and employee adoption of these tools can be challenging. How have you seen this with AI?
dictate emails, format them nicely, making forms, slides – AI has helped reduce the amount of time it takes for them to do that type of work. Trinchi: I run our IT team, and using paid and secure versions of tools like ChatGPT and Copilot has really ele- vated our team from a support and documentation stand- point. Historically, IT teams are not great at documenting
things. Now, when my team finishes a conversation with an AI tool that helps resolve an issue, I instruct them to document it. They can just ask the tool to write up a one- page summary, and then we add that to our knowledge base. That knowledge is extremely valuable. We have had people with tribal knowledge leave the company, and suddenly no one knows how to do their job anymore. AI gives us a way to capture more of that knowledge before it walks out the door. Blizzard: For Acme, our goal wasn’t necessarily AI; it was better decision-making for better outcomes. We cre- ated a simple maturity model. Level one is productivity. How do we get some of these tools in the hands of our employees to make them more productive? Level two is how do we integrate AI into some of the work? Third, is how do we get to autonomy? How do we educate where it’s making decisions for us? The human at the end has done a marathon for this decision. Here’s an example: Each day, he would spend about an hour and a half reviewing his batching report to iden- tify opportunities to consolidate shipments and save on freight costs, and send that information to our client rela- tions team We purchased a ChatGPT business license, configured the settings, gave him about 20 minutes of training, and provided the tool with a persona and specific instructions. When he uploaded the report to ChatGPT, it produced a full analysis within minutes. Not only did we save him an hour and a half to two hours a day, but we also uncovered opportunities he had missed. The tool identified addition- al freight savings because it could see patterns that were difficult to catch manually. That became our first quick win. Zlatic: That’s a great example of something people can take action on this week. I have heard that the best way to 22 June 29, 2026
Trinchi: It all comes down to the governance of it. If we do not provide an approved AI tool or a clear AI policy, people will go off and do their own thing, which is scary. There has to be a level of control. I am an IT person, so I am naturally a control freak. The job requires it. You need controls around your data and the use of AI. We are trying to limit the exposure of which AI tools are available. Zlatic: Everything starts with data, and consolidating it all is no small undertaking. Jason, what are your thoughts or best practices from President Container on bringing data together so you can get more value from it? Alvarez: There has to be one source of truth. You need a data scientist, a data engineer, or someone who can help organize your governance. Operations teams know how to execute, but someone has to organize those concepts across the business. That’s where I would start: establish a source of truth, organize it with rules, and align on the tools. Blizzard: I agree. If you can centralize your data, that’s a great start. Most of us have many different systems with data sitting in different places, and you need to know where that data lives. If you need help and do not have an internal IT team, partner with someone who can per- form a current-state data analysis. From there, identify the business problems you are trying to solve. You also need a full picture of your key business metrics, whether that is waste, on-time delivery, OEE, or something else. There may be critical data sitting in a system you do not even realize is contributing to that metric. Zlatic: A fear many plants have is investing heavily in AI without seeing significant results. What ROI have you seen from your projects? Trinchi: The ChatGPT licenses are worth their weight in gold. They have significantly augmented my team to the
CONTINUED ON PAGE 24
www.boardconvertingnews.com
Made with FlippingBook interactive PDF creator