From Playing With AI to Putting It to Work: A Practical Guide for Supply Chain and IT Leaders

[Transcript]

[00:00:00] Jan Griffiths: This is the Auto Supply Chain Champions Podcast. We are on a mission to bring you real conversations with the leaders who are transforming supply chains in the automotive sector. These leaders are true champions of manufacturing, and we're here to share their stories. I'm Jan Griffiths, your host and producer, and I'm joined by my co-host, Tom Roberts, Vice President of Strategic Industry Development at QAD.

[00:00:30] Tom Roberts: Great to be here, Jan. What I see every day is simple: manufacturers don't have a data problem, they've got an execution problem. This show is about how artificial intelligence, systems of action, and empowered teams can help close that gap.

[00:00:45] Jan Griffiths: Let's get into it. This podcast is powered by QAD RedZone.

Hello, and welcome to another episode of the Auto Supply Chain Champions Podcast. Let's check in with my co-host, Tom Roberts. Hey, Tom, how you doing?

[00:01:00] Tom Roberts: Hi, Jan. We're talking about AI today so I'm pretty excited, so let's go. Let's get into it.

[00:01:06] Jan Griffiths: I know, I know. You know what, Tom? I've been using AI more and more and more, and I can't even begin to tell you, you know, Claude is my new best friend now.

[00:01:19] Tom Roberts: Oh yes.

[00:01:20] Jan Griffiths: I wanna tell you this, I started off with ChatGPT, like a lot of people, and then Cheryl Thompson got me into Claude, and oh my gosh did my life change. And that's why I am thrilled today that we bring onto the show, Cheryl Thompson. Cheryl is the founder of the Cheryl Thompson Advisory and AI Adoption Advisory Practice because, Tom, many of us, you know, we play around with it, but we don't really know what we're doing.

I'm gonna be honest, I'm gonna fall on the sword right here and just say, I don't really know what I'm doing. So, I play with it and then I get frustrated and then I come back with it. We've got companies like QAD that manage the enterprise aspect of AI, but it's the average person. I don't know what to do. Who do you go for help? Well, we got Cheryl. So Cheryl, welcome back to the mic.

[00:02:22] Cheryl Thompson: Thank you. Thank you for having me, Jan. I'm so excited to be here. And hello Tom, I'm pleased to be on this podcast with you. I've been listening to some of your episodes and so important to talk about supply chain in the automotive industry these days.

[00:02:39] Jan Griffiths: Yeah.

[00:02:40] Tom Roberts: Thank you for that. I mean, it's one of those things, Cheryl, for me, where a lot of my job responsibility is working on artificial intelligence for manufacturers, but I think sometimes it's a physician heal thyself sort of thing, where I don't use it in my personal life as much as I should and could. So, I'm really interested in hearing from you, you know, where we can start to apply those things, not only in the business environment, but in things that can help us in our own activities as well.

[00:03:07] Cheryl Thompson: Absolutely, it's such a great place to start.

[00:03:11] Jan Griffiths: Now, Cheryl, I'm gonna take one of your lines away from you because I think it's so impressive. You've got over 1400 hours in learning and honing your skills on AI, did I get that right?

[00:03:26] Cheryl Thompson: Yeah. Yeah. It's around there. I am obsessed. I've gone down the rabbit hole, for sure. I started just about a year ago. I ended up coming across a podcast and they were talking about an AI bootcamp and I had been playing around with ChatGPT myself, and not getting great results and I thought, I'm gonna go to this bootcamp, it's $47 what do I have to lose? And they talked about a case for a tax strategy that really interested me. So, I thought, oh, I'm gonna go to this bootcamp, and after the three day bootcamp, I was hooked because I could see how I could apply it to my business in so many ways. And a year later, I've gone through their 12-week training program twice. I'm about to go through it a third time because, as we know, things keep changing, and I'm just really excited to keep going. So, yes, I've gone down the rabbit hole. It's been a journey for sure.

[00:04:20] Jan Griffiths: But you saw the need, you identified the need.

[00:04:24] Cheryl Thompson: Yeah.

[00:04:24] Jan Griffiths: Tom and I are starting to experience it ourselves, right? We know how to play around with it, but to really make it useful. So, tell us a little bit about that thought process and you know, and now you're all in.

[00:04:37] Cheryl Thompson: Yes. Yes. Well, I started to see the use cases both professionally for me as a small business owner and then me personally, like, how can it help me at home? How can it help me resolve conflict? How can it help me reduce overwhelm? And so, I think using it in your personal life gives you that confidence to start, you know, knowing how to push back on AI because AI will lie to you all day long if you're not challenging it properly, right? And then, you can start to use it in your business or, you know, on the job depending on where you sit.

And for me, as a small business owner, I'm a solopreneur right now, a one woman band, right? And it helped me with things I had only dreamed of doing, like technology is just moving so fast, so there's a lot of tools I wanted to use like Notion is something. It's like Google Sheets on steroids to organize all of your material. And about two, three years ago, I wanted to use that, but myself and the team couldn't figure it out. With AI, a couple of things, right now it's got an AI assistant in there so it can help you.

Now, Claude connects to it, but you can also just talk to AI and say, treat me like I'm in sixth grade, walk me through this step by step and you can take screenshots and put it in there and it will guide you through. So I feel like I have a tech coach in my pocket.

[00:06:04] Tom Roberts: Cheryl, what is the one prompt or one personal use prompt that always blows people's minds when you talk to 'em about it? Well, what's the one thing that really just has the wow factor to it?

[00:06:14] Cheryl Thompson: Oh my gosh, there's so many different ones. And this is gonna sound simple, it's meal planning. People love the meal planning.

[00:06:21] Tom Roberts: Okay.

[00:06:23] Cheryl Thompson: Right? Especially around the holidays, there's so many activities going on, thanksgiving through Christmas. So, if you are someone that is hosting or if you have to bring a dish, AI can help you with recipes, ingredients, grocery list, and it will organize by aisle in your grocery store, which is really cool.

[00:06:44] Tom Roberts: Oh, my word.

[00:06:45] Cheryl Thompson: I was using it, I know, it's so crazy, and I was using it for thanksgiving.

[00:06:50] Tom Roberts: So you mean it'll organize your grocery list literally by how it's mapped out in your grocery store, so you know.

[00:06:57] Cheryl Thompson: Yeah, you just tell it which grocery store.

[00:07:00] Tom Roberts: Wow.

[00:07:00] Cheryl Thompson: I shop at Meijer's, so it gives me the list according to Meijer.

[00:07:03] Tom Roberts: I did not know that. That is really pretty cool.

[00:07:05] Cheryl Thompson: It is very cool. And then, I was making thanksgiving dinner, the Lions game was on, was it Lions on Thanksgiving day? And so, everyone was watching the football game, and so I'm talking to ChatGPT, I put it into voice mode because my hands are, you know, I'm working in the kitchen, and I'm trying to work with the timing. So, it was walking me step by step. Okay, the game's over, now you gotta start to put the rolls in the oven and take this out and do this. It was, it's just like my assistant.

[00:07:33] Tom Roberts: Awesome.

[00:07:34] Cheryl Thompson: I love it. Yeah, I don't even use recipes anymore because I'll just go to ChatGPT is my favorite for cooking, like last night I made a phenomenal chicken pot pie and it was so good. Just walking me through.

[00:07:48] Tom Roberts: Comfort food.

[00:07:48] Cheryl Thompson: Yeah. Yeah. It was amazing. It turned out phenomenal.

[00:07:52] Jan Griffiths: Who would've thought this is coming? AI and food, who would've thought, but here we are.

[00:07:57] Cheryl Thompson: Hey, everyone's gotta eat, right?

[00:07:58] Jan Griffiths: Here we are. Yeah.

[00:07:59] Tom Roberts: That's right.

Yeah. And Cheryl, so you you talk about AI evolving from an assistant into something more, and this is kind of what you are referring to now.

[00:08:07] Cheryl Thompson: Yeah.

[00:08:08] Tom Roberts: Walk us through more of that. What is the difference between a custom GPT, a Claude project, a Gemini gem in plain english, for me.

[00:08:17] Cheryl Thompson: Yes, yes.

[00:08:18] Tom Roberts: Sixth grade level. Walk us through those differences.

[00:08:20] Cheryl Thompson: Yeah, so I would say an AI assistant is using a prompt. And I like to call them repeatable prompts. My structure is RCRQ, so you have to give it a role. Who do you want AI to be? Because let's face it, the Internet's full of all kinds of information and the role helps you hone things in.

So, if I'm trying to set a boundary, I may say, you are a Henry Cloud author of the book Boundaries, and then you give it a little context. It needs to know about what pain or struggle or problem are you facing? And then, you go on to the request, give it a little bit of guidelines, constraints, what do you want the output to look like? And then, I always end with, please ask me any clarifying questions.

So, that's using AI as an assistant, right? It's helping you. Then, if we go beyond that, we think about it as a specialist, we are now talking about creating a Custom GPT or a Claude project or a Gemini Gem, and they're all the same things. So, it is a specialized tool that has a set of custom instructions that go a little bit deeper than a prompt.

And then, in the backend you have a knowledge base, where you can put files in there. So, one of the specialists that I have in my business is my copywriter and my marketing assistant. So, it learned my voice. I have my voice profile in the backend as in one of those knowledge base files. I have past customer testimonials in there. I'll have my past social media top performing posts in there. I've got my website copy in there, so it knows a lot about me. So, you set that up and then you train it.

So, in this case, I would say give me 10 blog ideas. And then, I'll pick one and then I'll say, write me a blog and then I read that blog and then I'll say, wait a minute, I would never say this. And then, I keep working with it to amend the custom instructions to get it like I want, so that would be a specialist.

Now, and you can do that in Copilot, that would be an agent. You can do it in Gemini, it would be a Gem. In Claude, it's a project. And then, chatGPT, it's a Custom GPT. Now, AI is evolving so fast that you can also have employees, so that would be like Agentic AI agents. So, chatGPT just came out with their workspace agents. I have not played around with that yet, but I have played around with them extensively in Claude.

So, an AI employee would be, for example, I am looking to get more speaking engagements. So every Monday morning, I have a scheduled task that runs and I will say, please look for speaking opportunities. It will give me the top 10. It will score them. It will say, these are the three you should apply to. It will tell me the email addresses or the websites I need to go to. And then, it will get, it will draft outreach messaging for me. I could take that a step further and have it send the email, but I still want the human in the loop and I wanna see it before it gets sent.

So, I have about nine of those automated tasks running and you know, you could think of them as little mini employees. So, we've got AI as an assistant with the prompting. We've got the specialist with a custom GPT or a Claude project. And then, we've got like an AI employee, which I would say in the last three months has really come to light.

[00:11:51] Tom Roberts: Wow. Awesome.

[00:11:54] Jan Griffiths: Where do people struggle, Cheryl, the most? People who are starting with AI and moving from just playing around with it to actually making it useful? What are some of the maybe more of the psychological things that hold them back.

[00:12:07] Cheryl Thompson: Well, I think there's a few things. I think just using it like a search engine is one thing, right? We have to be very specific with it. Like, who do we want it to be? What do we want it to do? Also, I don't think a lot of people know you can set up system instructions, so you're telling AI how you want it to talk to you.

So, you know, you can have AI interview you for that and then put that in your system instruction, so I think people are missing that. I think that people may go ahead and have AI do something for them, and they don't like the result, and then they just think, oh, this is, this is trash, right? And think about that person that you worked with that was very loud and very confident, but they didn't really know what they were talking about. AI does that, too.

[00:12:57] Jan Griffiths: I know a few. I know a few.

[00:12:59] Tom Roberts: 100% confidence on AI.

[00:13:00] Cheryl Thompson: Right? Yeah. Yeah. AI will come back and it, it will sound so true. And then you look at it and you're like, wait a minute, that's not true. Where did you get that? And it'll say, oh, you know, I made that up because AI wants to please you. It wants to be helpful. So, you have to really push back on it. And in my system instructions, I'll say, push back on me. Tell me what I don't wanna hear. So, I think that that's some of the ways people get tripped up.

I think people are worried it's gonna take their job, and I always say, AI is not gonna take your job. Someone that knows AI is going to take your job. We have to start thinking about these roles like a year from now, five years from now, what is this role gonna look like? Because I think roles are going to evolve as agentic AI comes more and more into the picture, and we've got these little assistants working for us.

[00:13:49] Tom Roberts: Yeah. One of the things that we talk about at work here is, if we automate, say the RFQ process, right? For suppliers. And you know, maybe you were doing three or four potential suitors for something. You were going out, setting up the RFQ, 'cause there's just so much administrative work.

[00:14:05] Cheryl Thompson: Right.

[00:14:06] Tom Roberts: Are people prepared to do 15? Because AI is actually reducing the process cycle time down to maybe 30% of what it was, or 25%. So, are people really in their minds ready to say, I now don't do three or four. I can look at 15 different potential suppliers, so that's the biggest thing is Parkinson's law I think they call it.

[00:14:29] Cheryl Thompson: Yes.

[00:14:30] Tom Roberts: Work expands to fit the time alloted.

[00:14:32] Cheryl Thompson: Right. I love that.

[00:14:34] Tom Roberts: Are people ready to say, okay, I've saved all this time, great. Now, what do I do?

[00:14:39] Jan Griffiths: Yeah. Now what?

[00:14:41] Tom Roberts: Now you can handle more or you can do a strategic thought or, you know, whatever else you need to do. But you need to fill that time, and we warn people, we warn our customers that you are going to have this time, you need to know what you're gonna do with it first before you implement the software. Otherwise, often don't come up and say, hey, I've got half my day free now can you gimme a bunch more work? It doesn't uniformly happen like that all the time, so.

[00:15:07] Cheryl Thompson: Tom, this is so important. I'm so glad you brought this up because when I first got into AI, I got into this trap of producing, producing, producing, right? I can do like five things at a time, and then I found that I was just filling my plate with more things and I was really feeling a little bit scattered because I wasn't building in those mental breaks, right? So we have to be a little bit careful. We can't just overwhelm ourselves with, hey, I can do more, doesn't mean I should do more.

And one thing I've started to hear from people is their managers or their supervisors are expecting more. Hey, just use AI, right? I can now pile on more work because AI is gonna help you, and I think we need to be really careful about that. And I love what you said, like, be specific, like know what you're gonna do with that extra time. So important.

[00:16:01] Tom Roberts: Right.

[00:16:03] Jan Griffiths: I never thought about that, Tom. So, you seriously, you have to talk to clients about that, right? You have say, look, we're gonna save you so much time. You're not gonna know what to do with it.

[00:16:12] Tom Roberts: Yeah, because when you implement it, and again, I'm not indicting anybody, but you know workers tend to, you know, oh, I've got this great tool help me out. Good. I'm using it, and now what do I do? Some of them will come forward and say, I want more, or I have more time. I can do more things. Or I can spend some time. Jan, I can spend more time negotiating. If I don't have as much time doing the administrative work of sourcing, for example, maybe I can spend more time in negotiation or supplier collaboration, those kinds of things.

So yeah, we definitely tell people you are going to have a process cycle reduction time for the work you have today. You've gotta figure out what you're gonna do with that. And we introduce things like more strategic thought or, there could be, again, adding additional potential operational items in there.

Sometimes, a company might not want to have a particular person spend time in strategic thought because maybe that's not their asset, right? Not their thing that they do. So maybe they need to do some other headcount optimization thing. So, it's very complex. What we're noticing is these are great tools, but they are happening so fast that they can completely change the organizational psyche of a company very quickly.

[00:17:29] Cheryl Thompson: Yeah. And to your point, I think we need to do some visioning about how these roles are going to evolve, like you said. Yeah. There was some data from Asana that said we only spend 27% of our day on the work we were hired to do, because so much of it is administrative, right? So, if we can free up people with that administrative, they get to spend more time on what they were actually hired to do.

I remember when I was an engineer, was I really doing engineer work? No, I was doing so much administrative work and project management. And then, you were talking about negotiation. I mean, that's a great use case for AI. Just, you know, spin up a little prompt to, or a custom GPT or agent whatever to role play with you on that negotiation. Like, push back on me, prepare me for this negotiation. I think that's a great use case.

[00:18:20] Jan Griffiths: And you know, it's taken me back a few years now I'm thinking about my time as a buyer a long time ago, and then run in procurement organizations. And it was always this struggle because you want the buyers to be out with suppliers using their human skills that AI can't replicate: building relationships with suppliers, understanding what new technology suppliers have, looking at suppliers.

I mean, it is using those human skills to build trust, to build that relationship, to build the highway to allow things to run at speed with AI. That, to me, is a massive, massive shift and one that I love. And as I look back, I spent so much time on spreadsheets, which I hate and I'm no good at by the way, but on spreadsheets, crunching, you know, quote comparisons, quote details, cost breakdowns, metal adjustments, all of these things that you gotta do. Approving POs. Looking at purchase requisitions. I mean, I could go on forever, right? All these things you can get an AI agent to do, whether you do it in your system as an enterprise level, the kind of work that Tom does, or whether you do it on an individual level, which is more of the work that you're focused on, Cheryl.

[00:19:37] Cheryl Thompson: Yeah.

[00:19:37] Jan Griffiths: But you can use AI to do that, so it's a huge opportunity for us to move forward in the function.

[00:19:43] Cheryl Thompson: Yeah, and I'm so glad you said that because we have to remember the human in this AI world. And so, to Tom's point of, what are you gonna do with that time? Spending more time face-to-face, building those relationships are so important. And Excel, you don't need to be an Excel whiz anymore because AI is that Excel whiz. Claude has got a phenomenal AI connector with Excel. And Microsoft has an Excel agent that's pretty good as well.

So, it's those things that you hate to do and I think that's where people should start. It's like, what is draining your energy? What are those things that you do repetitively day to day that are administrative? And then just talk to AI, say, hey, this is what's on my plate. Like, I like to do a Monday morning dump, like everything that's on my mind, all the things I'm concerned about, the things that I've been procrastinating about. And then, I have a prompt, like you are David Allen, the creator of Getting Things Done, and then help me organize my day. Help me with the Eisenhower Matrix, important and urgent and all of that. Like, help me plan this out, and it just frees up all that mental bandwidth and the things that you've been putting off.

Now, you've got a plan for the difficult conversations that you are maybe dreading. AI can help you with that. I just helped my son with a negotiation at work with his salary, and he got like a 24% raise, but he got the confidence to.

[00:21:14] Tom Roberts: I need to talk to Cheryl.

[00:21:15] Cheryl Thompson: Like stand up for himself.

[00:21:16] Jan Griffiths: Oh my God. Everybody's gonna be calling you now, Cheryl.

[00:21:20] Tom Roberts: That's fantastic.

[00:21:23] Jan Griffiths: Yeah, that is great. Now, what advice, Cheryl, would you give to the average person working? Our audience is primarily supply chain and IT, so we cover traditional supply chain, purchasing, IT, and operations, so that's the audience that you're speaking to. Give them some advice. Let's say they just started playing around a little bit with chatGPT, and they're ready to move to the next level. What are three things they should do?

[00:21:50] Cheryl Thompson: Well, number one, I think that they should look at their to-do list, everything that's on their to-do list, and then just have a conversation with AI. Say, hey, here's everything on my to-do list, how can AI help me? Just start there.

And if it is, hey, I've got all of these supplier bids I need to evaluate. Maybe you can design a prompt to help you design the evaluation criteria that you're going to assess everybody against. Or if you've got a list of suppliers that you need to assess, like who are the ones that are at risk? Who are the ones that are top performers? You can have it organize that.

We talked about the negotiation one. I'm actually doing a learning lab next Tuesday for supply procurement professionals. And some of the prompts that we're gonna be looking at is, you know, drafting that supplier communication email, the follow up, all of that. An RFQ and documentation builder.

And you could start within your own team to use these repeatable prompts to set up standards, like, this is the way our team is going to work, you know, our documentation is going to have these in key important elements that would be another one. It's like talking to your team.

And then, start to use it in your personal life. I just used it to build a workout plan. I got injured over a year ago. I was doing a deadlift and I ruptured my L4 and L5 disc and was impinging on my S1. Yeah, it was so painful. I was like flat on my back for 10 weeks. And it took me really some time to get back active. And then, I injured my rib and so I was able to communicate with AI, here are the physical limitations I'm dealing with and it built me a workout plan with the workout people I like to follow.

It found some places to walk in my neighborhood that have nice walking paths and I built it in Claude and it's like an interactive website now that I have that kind of keeps me on track. So, start playing with it in your personal life.

I'm helping somebody with a job search right now and I was able to have Claude help me with what should a LinkedIn profile look like? What should the headline look like? What is the standard resume format you should follow for applicant tracking systems? And then, I had it crawl, you know, their LinkedIn profile and then, you know, based on that, which jobs should this person apply for? And then, I was able to make that into an interactive website so that they could see all of the jobs available, actually link to the posting so that they could apply, and then give them networking tips on who they should network with and what their outreach messaging could be.

So, the sky's the limit. The only limit is your imagination. So start playing around with it. I mean, whatever tool you have, many of these, all of these tools are free. Just to start playing around with, see which one you like. I love Perplexity for research 'cause it gives you the sources and then you can say, where do you want it to search?

So, if I'm getting ready to do a workshop, I will have it search not only the web, but social media. And I'll say, what are people saying negatively about this topic? Because I wanna know what are the complaints going to be before I get there. So, Perplexity for research. Gemini for image generation. Nano Bananas, great, but ChatGPT is good and Claude's coming along as well, but find the tool that you really like. It's like Mac or PC, which tool do you like, and then go all in with it.

[00:25:33] Jan Griffiths: I absolutely love that. Now, you run a lot of workshops, Cheryl, and they're focused for job type, for function, for different companies, and we'll drop a link in the show notes so people can go through that and see if they wanna sign up for any of those. But I have learned so much already in this session of you. Tom?

[00:25:54] Tom Roberts: Absolutely. When I get into Claude and figure some things.

[00:25:59] Cheryl Thompson: Yes, yes.

[00:26:00] Jan Griffiths: It's been great and you know, Cheryl, thank you for putting all these hours to do the work that really not a lot of people wanna do. I do not wanna spend 1400 hours learning AI, but to spend a couple hours with you on a workshop, giving me the benefit of your knowledge, yeah, I'm in on that.

[00:26:19] Tom Roberts: I'm just listening to you, Cheryl. It's warp speed. I mean, you literally are taking people, you know, much farther than they were, so it's great.

[00:26:26] Cheryl Thompson: Yeah. Yeah. I love it. I'm obsessed.

[00:26:31] Jan Griffiths: Well, thank you for joining us

[00:26:33] Cheryl Thompson: Thank you for having me.

[00:26:36] Jan Griffiths: We wanna hear from you, our listener. Tell us what are your challenges right now? What conversations do you want to hear across the airwaves on this podcast? Drop us a comment on our podcast website. The link is in the show notes.