How One Manufacturer Took AI to Prime Time – And It Worked

According to Tim Dickson, the time for “playing around” with AI was last year. This year, the stakes are higher.

Dickson, Chief Digital and Information Officer (CDIO) for Regal Rexnord, believes it’s time manufacturers start proving the value. And for Regal Rexnord, a global power transmission and motion control company with more than 50 brands, that means getting past the low hanging fruit and on to the game changers. We recently spoke with Dickson for some detail on what that process looks like and why Regal Rexnord is committed to moving beyond basic machine learning.

(Interview has been edited for length and clarity).

Manufacturing.net: First I want to talk a little bit about the goals for AI. I think there are a lot of organizations out there that are really looking at this in the abstract; they want to use AI, but don't really know how. So, in your opinion, how should they be exploring this and who can they be looking for guidance?

Tim Dickson: I was brought into Regal Rexnord as a CDIO to use digital technologies, whether those be AI, IoT, cloud, e-commerce, to try to figure out if they could provide a 50- to 60-year-old industrial manufacturing company growth. And so the value of AI is talked about quite a bit; kind of a taboo term in some companies. But we are really on a mission, and I'm personally on a mission to show how digital technologies like AI can actually provide revenue growth for a company.

Mnet: Do you have any examples of exciting applications that you could tell us about?

TD: I truly believe that in order to get value from AI, you have to go through an AI journey. We started off very basic, with traditional industrial manufacturing AI use cases like cross-sell, upsell, demand and sales forecasting, online product recommendations and predictive analytics –  things that machine learning models have been delivering in industrial manufacturing companies for quite some time. 

That has generated a significant amount of benefit for the company  – one, to show that an IT organization can transform to a new digital mindset and deliver and incubate AI, but [also] showing that it can improve bottom line and top line value was huge. So that was just demonstrating to the company that we could do AI.

Tim Dickson, CDIO, Regal RexnordRegal RexnordThen we moved on to generative AI, and the transition was: Could you do that in a setting where you're now interacting with either an external customer or an internal team member? Could you change the behavior from a customer calling into a call center to now interact with you via chatbot? Could you change the interaction from an internal employee [asking a question to a chatbot and getting an immediate answer]? 

So proving that gen AI can then be productive, that people would want to interact both internally and externally through a chat bot interface that was intelligent, that was very responsive, that was content-rich and provided videos, pictures and images were our second phase of our transformation. We launched a generative AI chat bot on the RegalRexnord.com website called RXXy, and we never would've thought the usage and the attention would be so high from a customer set.

Legacy industrial distributors have been around for a long, long time that formally interacted with sales and customer service over the phone now can get their questions answered via a chat bot interface that has yielded a ton of usage, a ton of customer satisfaction, a ton of answers [to questions].

We've been able to save a significant amount of what I refer to as internal productivity through our internal chat bot for employees and team members: How many vacation days do I have left? What's my benefit policy? Who is Tim Dickson in this company? I have a meeting with so-and-so I'd like to prep for him. What does he do? What does he like? What's his profiling?

The second phase of our journey, now we're on to agentic AI. This is a little bit tougher, but I do believe this is where the most savings or opportunity can be. Now decisions and actions are being taken as a result of the data and the insights being delivered as a result of the model. So can an agentic AI agent interface with your customers for a certain number of products or a certain number of regions or a certain number of customer distributors? Could an agentic AI bot make decisions and reply to emails from customers based on a certain topic or responses? Could an agentic AI bot respond to a customer's quote, respond to a customer's query with a valid price with their discount? So there's a whole new world of agentic AI that we're currently in the process of piloting that is a bit more exciting and potentially game changing.

Mnet: One other thing I wanted to touch on here is around risk and risk avoidance. I'm sure that there are lots of stops and starts with an evolving technology like AI. Is there anything that you could tell us about recommended strategies there for avoiding some of that risk or maybe course correcting there?

TD: I've got a few examples of those. Failing fast especially, but before I get into the cultural side … [the goal is] promoting a risk-free environment and not having risk aversion, because some legacy IT organizations do have a significant amount of risk aversion. 

What I like to tell my teams is, you’ve got to do the basics and the foundational things. And then you’ve got to promote a culture and a mindset. And the basic and foundational things are, if you're doing anything with AI, you need a documented AI policy. And I chose to have my cybersecurity team document my AI policy, which is really about copyright, accuracy, bias, privacy and then the training required to even trust that you're going to be given a data set from the company and you have some training or certification, you know what you're doing and you can actually leverage AI.

Regal Rexnord's digital assistant, RXXy.Regal RexnordSo you have to have some sort of policy. And I'll tell you with the last year and a half with regenerative and agentic AI, that policy has had to get updated, no doubt, because the early models are a lot different than the models as it lay with all these bots. So you have to have the foundation in place. If you ever launch anything with AI, you have to have an AI governance team or an AI governance model. Let's take in some new AI ideas and use cases, but it's also risk audit security, legal data retention, data privacy, those sorts of things so that the governing body of the organization knows what you're working on and is there to provide guidance to prevent any hallucinations or bias or things of that nature from an audit, cybersecurity, legal and risk perspective.

If you're doing anything with AI, certainly on a multi-billion dollar revenue generating enterprise, you have to have those foundations in place. So if you have those, then it's really about the culture. As a leader, do you allow your team to create potential new solutions that have never been proven before that might provide value in the organization? The way that I do that is we have events in the form of hackathons … jam sessions; I hosted the company's first generative AI leadership summit in December where we brought in generative AI platform providers where they demoed their solutions. And my business partners and my IT organization team members got to learn and ask questions all at the same time. And that generated this whole long pipeline of requests that we've implemented or are in the process of implementing right now.

Mnet: Is there anything that we have not covered yet that's very important to you about this conversation?

TD: There aren't a lot of people talking about value. I personally think last year was the year of playing around with AI; I think 2025 has to be about proving the value. Is that model that you're building somehow generating insights that are contributing to sales or a financial decision that somebody else is making as a result, and then somehow enabling some top line or bottom line value down the road? 

Can you actually demonstrate that people are making different, more effective decisions as a result of the insights that you're delivering? Or are you just using the AI as a more efficient way of doing your job, but it's not necessarily game changing or hugely impacting it in a positive or negative way? So I'm trying to prove that AI – particularly generative AI – is actually providing growth and demonstrating that growth in any number of ways.

I was actually hired to drive digital transformation at this legacy organization. To be told that you can innovate as long as you're innovating with an actual business purpose and value is a dream come true. That is something that most CIOs do not have.