← The Amazing Intern - Turning Your A.I. Curiosity Into Confidence
The Amazing Intern - Turning Your A.I. Curiosity Into Confidence
Ep 79 - Getting Your Team to Use ChatGPT: Real Lessons from Bill Gross of the Nuclear Energy Institute
00:26:41
n this episode, Cary talks with Bill Gross, CIO of the Nuclear Energy Institute (NEI). Bill shares the story of how his CEO tasked him with preparing their 100+ person organization for an AI-powered future. With a background in computer science but...
Loading summary
Transcript37 lines
- [00:00]
T Mobile Advertiser
Everyone's loving family Freedom from T Mobile we'll pay off four phones up to $3200 and give you four free phones all on America's largest 5G network. Visit your local T Mobile location or learn more@t mobile.com FamilyFreedom up to $800 per line via virtual prepaid card typically takes 15 days. Free phones via 24 monthly bill credits with finance agreement eg Apple iPhone16128GB8999 Eligible trade in eg iPhone11 Pro for well qualified credits end and balance due if you pay off early or cancel. Contact T Mobile support for this podcast and the following message comes from America's Navy the Navy offers new graduates hands.
- [00:36]
Kerry Weston
On training and experience in careers like.
- [00:38]
T Mobile Advertiser
Computer science, aviation and medicine plus education and sign on bonuses. Parents help your grads start their career today@navy.com.
- [00:51]
Bill Gross
We'Re located in Washington, DC. We have about 100 to 110 staff. I took over the role of our CIO in January of this year. At about two weeks into it, my CEO comes down and says AI is going to change the world and I want any IDE to be ready for it. So your job is to upscale our entire company.
- [01:15]
Kerry Weston
Hey there. Welcome to the ChatGPT experiment. This is a podcast designed to help you better understand just what ChatGPT is, what it can do, and how it might be able to benefit you in a personal or professional manner. My name is Kerry Weston. I'm your host. Welcome to the show. Hey, today I've got a conversation with Bill Gross, who is a listener. Bill reached out to me from the Nuclear Energy Institute in Washington, dc. He had feedback on an episode. But you know what? It didn't take long. As you heard in the clips at the very beginning. It didn't take long for me to understand that he's had some experiences in his own company that I think you all could benefit from greatly. Especially if you're looking at your own company and saying, I, I know I've got to get AI into the company. I know we've got to have more people experienced and understanding and curious and testing. How do I do that? Well, Bill was charged with that, as you heard, merely two weeks into a position at the company and he put some, some tasks together, put some tips together and put some processes together. He's going to share with you that regardless of whether you have 100 employees like his organization or just five, the ability of getting people to adapt, to be curious, to share, you're going to hear Some of his challenges. You're going to hear him say, I wish I had done this. This is a lesson learned. Or I did this early and it benefited me greatly. Right, so there's some nuggets here that Bill's going to share that I think is extremely, extremely valuable. Bill's a listener of the show and he reached out for something completely different, like I mentioned. But if you have a nugget, if you have an experience, if you have a story that you think would be worth sharing, let's learn together. Right? ChatGPT, experiment.com is the website. You can find how to reach me there. Also, as I mentioned last time, I've got a page that I'm putting together now with guides, tips, prompts, and resources you'll find in the main menu under training. You'll also find ways of getting a hold of me. If you want to do an individual training or a business training, or if you've got a workshop coming up, you can, you can get a hold of me there. Hope everything is going well in your world. I love it when we get to share nuggets, share experiences, share best practices for the good of everyone. So this is a good one. And if you have a thought, a goal, a challenge of getting this into an organization that you work with, hopefully Bill's got a couple nuggets and lessons that will benefit you. Okay, so let's get you to the conversation I had with Bill. And as I always say, stay curious. Okay, talk soon. Well, thanks for reaching out.
- [03:43]
Bill Gross
No problem. No problem. Love the show. I don't. I, you know, honestly, I don't recall kind of how I found you, but probably heard your show referenced on another show and added you to the queue about three weeks ago and have probably gone back and listened to a third of your shows, maybe just to paint a little bit of the picture. I took over the role of our as any CIO in January of this year, and about two weeks into it, my CEO comes down and says, it's going to change the world. And. And I want any eye to be ready for it. So your job is to upscale our entire company at that point. You know, I have a background in computer science, and in the 90s when I was in undergrad, you know, we were studying neural networks and stuff like that. And, you know, that was the picture I had of AI was maybe 30 years old and a lot had changed, but my experience was tainted by those early days. So I'm like, okay, well, what is this AI thing? You know, January of this year, I still had no experience with it, so created ChatGPT account and probably 15 minutes into it realized this is monumental, what is about to happen. And since then it has been the slow road of trying to take an organization that is just busy to the core and figuring out how to get them to carve out just enough time to get that aha moment, you know, which is one of the. One of the things I really love about your show is it's how do you get started at ground zero, which I think is really attractive. So.
- [05:21]
Kerry Weston
Well, Bill, could you. You mentioned an ei. Could you just explain what that is?
- [05:25]
Bill Gross
Sure. The Nuclear Energy Institute is a public policy organization. We're located in Washington, D.C. we have about 100 to 110 staff. And, you know, we promote the beneficial uses of commercial nuclear power.
- [05:40]
Kerry Weston
Okay. You had reached out. I think there was probably an episode that triggered you reaching out, something around projects, right?
- [05:49]
Bill Gross
That's right. I think it was relatively recent on the show. I think you had been articulating that you were principally a custom GPT user, hadn't spent much time exploring projects. I think you characterized them as really just a container, like files in a folder. I think there's a lot more use cases there that I think really are an unlock. At least they have been for me personally.
- [06:17]
Kerry Weston
Yeah. Yeah. So that was really interesting. I appreciate that. And yeah, I've only dabbled. I found a little bit more value to it. But I love the fact that you listened and said there's more to this and wanted to reach out and I'm glad we're having this conversation. So I thought record it for the show in hopes that someone can learn something as well. So what are you thinking? Start me off with what your first thought of. When you said. When I had mentioned containers and you said there are more value, give me some examples perhaps of the value.
- [06:48]
Bill Gross
One of the differences that I think is at that intersection between custom GPTs and your regular chat thread is within projects. You can have context files at the project level and every conversation you have within that project can reference those project files. It's very similar to knowledge files you might upload to a custom GPT so that every time you exercise that GPT, it has those knowledge files. But I think the unlock is when you use a custom GPT. It's once and done, you run the GPT, you get the output and you do something with that. In my role, I find a tremendous number of use cases where I'm going to create a very lengthy set of threads within a Specific topic and I want the same context files to be shared across all of those threads. For example, when my CEO left the office, I created a project I dumped in a file that is high level overview of any our current organizational structure, our AI policy that we had at that time. Then I just started creating threads in there. How do I establish a training program? Should I establish a champion program? Each one of those had those knowledge files that it could reference. It started proposing different folks in the organization we should reach out with to create a supplementary steering committee or folks that might be rich targets to join our champion program. It had a lot of context that each thread within that project has knowledge of. I found that a really compelling use case.
- [08:36]
Kerry Weston
Yep, great. Is there another value that you've brought or you found inside projects besides that?
- [08:43]
Bill Gross
To me, that's one of the biggest ones. And the use cases are pretty lengthy. You did a show recently on using it as sort of a thinking partner or coach and that was like my second use case. I started a new job in January, whole new team that reports to me. So, you know, putting in a project file of the scope of my responsibilities, the scope of the people that report to me and just having sort of a single place that I could go and create threads maybe for each employee within that. As a project where, you know, my strategic goals for each of my employees is an element of that project, it can start to see synergies across the departments and whatnot. Yeah, I love it.
- [09:29]
Kerry Weston
What's one of the things that you're doing at AI now that is being aided by your ability to use Chat GPT or is there something that's becoming faster, more efficient or more valuable now that you've been able to kind of explore how to use whether it's projects or just a conversation in and of itself.
- [09:45]
Bill Gross
So we, we settled very early on as ChatGPT as our preferred AI model within NEI just to try to create a common baseline of what the, you know, consistency and what the tool is and how different folks are using it. And we're seeing various use cases across different departments. As an organization, we're really busy right now. So it's been. I wish I had more data coming in on specific use cases that people have, but I think as people discover these, it's sort of just enough of a time savings for them, but they don't have enough time to like, let's cheer how we've been using it. Definitely hear a lot on the content development side or ideating side. I know our technical and regulatory Services group uses it really extensively for policy analysis. So there's been many executive orders under the current administration that affect energy and nuclear in specific. So sort of digesting those, bouncing those off of policy priorities, potentially doing regulatory analysis. It's been a very useful tool for them. And you know, the scope of what it's able to process in terms of information versus a human is very, very large. One of our very earliest use cases in that group that sort of made the case for that department to really fall in love with it was the Nuclear Regulatory Commission that regulates commercial uses of nuclear power issued a revised guidance document. This document is about 800 pages long for a human to pull the prior rev and pull this rev and say, well, really what's changed is weeks of really reading and trying to understand it. And our lead in that group basically dumped them both into ChatGPT and said, Tell me the differences and how important are they? And it did it in about two minutes. And that's amazing time savings.
- [11:51]
Kerry Weston
So a couple of things you shared with me, which probably was a very eye opening directive, is, you know, being charged with bringing AI into the organization and making sure that it's understood and kind of implemented. What was it like to hear that and how did you go about it? Because I know that there's others listening that have been charged with something similar. They may not have 100 employees, they may have five. But how did you actually approach the fact that everyone kind of had to have a base level knowledge of what was going on and how to use a tool?
- [12:22]
Bill Gross
One of the very first things I did, my natural tendency is to learn as much as I possibly can. That's where I go by default. I started listening to, I don't know, 30 podcasts a week on AI and AI implementation. One of the first things that came up was you definitely need a steering committee and a policy. We had that that group had been established for some time. We had a well established policy. The next was develop a champion program, a subset of the staff that you can invest the time in to upskill them. And if you get a broad enough selection from different groups and different organizations, they will then take that knowledge, work with their peers in the group because they know the use cases that are relevant to their team or where the opportunities are within that team. So somebody in our finance group, somebody in our communications group, someone in our technical and regulatory services group. So we did that, we identified, we got about 25 volunteers, which I was really excited about about. I mean it's almost 25% of the organization, we did send them to some training and then and then kind of cut them loose and some, you know, they're all volunteers and some of them are really trying to find ways to work this responsibility into an already busy schedule. I think that's been effective. But you know, we call these folks our jedi. J E D A I kind of cute play on the Jedi term there. But they do need care and feeding, they need some encouragement, they need management support and they need to feel like it's worth their time to carve time out of their schedule to do that. So I think they're really the keys to success. But it's not a turnkey solution. You definitely have to invest in that inside your organization.
- [14:13]
Kerry Weston
What is the process for folks to actually share those successes or lessons learned with the group, be it the 25 or your, you know, 100 at large.
- [14:22]
Bill Gross
We've tried a couple of different things and I would be happy to hear other ideas on this. We do, we have, you know, we do everything hybrid because we do have some, you know, permanently remote staff or staff that might be working from home on a given day. So we do everything hybrid. We've done lunch and learns. So we've done a couple that are really sort of deep dives into just the ChatGPT interface. What are the features? What are the different models? What might you use them for? We covered tasks, scheduled tasks, which I know you did a show on. So we've done some basic upskilling like that. And then we really try to elicit from our JEDI examples use cases. People love to hear from their peers. We've tried to do as much of that as we can. I wish the pipeline of that data coming in was bigger. We started now doing office hours where people can just come and bring use cases and we'll try to help them through that. Sometimes ChatGPT is the answer to this. Sometimes other tools might be appropriate. Then every fall in late October, NEI does an organization wide training week. We bring the whole staff here, even your remote folks, and we spend a week on site and we're going to do half a day worth of deep dive into some AI topics. Very much looking forward to that. To me, my primary objective is to get somebody to put their fingers on the keyboard or to click that advanced voice mode button and talk to it because it does not take long of somebody interfacing with it before the light bulb goes off and it's sky's the limit after that.
- [16:04]
Kerry Weston
Yeah, I love that. And I'm sure that there are Others that have found ways of distributing and holding it as well. And if you're listening, you do, please share in because I think this is an issue. We don't want to create silos. We don't want to create an excited employee that goes off and does it by themselves and never brings that back. Right? The hunter, gatherer. We want them to come back and share. And so what we say is, just because the conversation. Just because you are new to the conversation doesn't mean the conversation is new. Because you're going to have, let's say, 100 people, you're going to have a subset of those that get really excited really quick, right? But then you're also going to have the majority that aren't, but at some point they're going to be interested. And wouldn't it be great if the person that is interested six months from now has the ability of tapping into the excitement and the lessons learned and the experiences that your early adapters did, Right? And so having a repository somewhere where you can share and have a library of either, be it tips or videos or something, lessons learned, where somebody on their own time can join the conversation, really get up to speed based on where others have been, that's the challenge that most people have, is where is that repository? Where does that volume live? Right? Where does that secret sauce live? So people can get up to speed on their own time. But I applaud you for kind of being open to go try and bring it back. And then it sounds to me like you've tried a few different ways now to get them to collaborate. But it does come down to having time and interest in doing so. Right. Of the others, you clearly can't force it, right?
- [17:37]
Bill Gross
That's right. As much as I want to. And we do have some members of our leadership team that have used it and are really sold. They understand the value and, you know, sort of finding that right balance. You know, there's a carrot and there's a stick, right? The carrot is we offer training, we offer. And the stick is now we're going to make this a part of our performance management program. Right? We want you to be identifying use cases or potentially identifying time savings. You know, I think we're, we're not there yet, and I think we need to be very, very judicious about when we make that decision. The other thing you mentioned that we're really spending some time on is as we hire new employees, how do we get them up to speed as quick as possible? Because they come in and some of the folks that we hire, I mean this industry is notoriously very, very risk averse. But we're AI forward. We want you to come and feel safe using these tools. So how do we get you in the door and up to speed as quickly as possible with AI? So we're working with our HR department to bolt something into our employee onboarding where we might have a one hour sit down with the staff, show them tools that are available and if we need to do some upskilling, we'll do it face to face with them.
- [18:56]
Kerry Weston
That's a great idea. And along the lines, if there's some sort of place there to put best practices or lessons learned kind of videos or tips for what? That would be a great onboarding exercise for a new employee or any employee really. But new employees tend to have more objective eyes coming in and maybe a little bit more curiosity to get up to speed. That's great. If somebody was listening and they were thinking of doing the same thing and having a company conversation or try to get folks inside the company kind of involved and realizing the value of this. Other than what you've shared already, are there any lessons learned or tips that you would give to a new person coming into this that would help them learn from where you've been?
- [19:36]
Bill Gross
Sure. Kerry. I really like that question. I think there are some things that I've done in the last seven months that I'm really proud of and I think have really worked. There are some other things that I wish I had done that might have made our journey a little bit more accelerated. One of the early things I did was about four weeks after my CEO came into the office with this mandate is I did a one hour training for her and her executive assistant. This was a hands on training using advanced voice Mode. She's a CEO. I'm not expecting her to type into ChatGPT. We had her installed the OpenAI client on her phone and I prepared a series of prompts that added slightly more info to make the case of how important context is. We did this training and she loved it and had me do it for entire executive leadership team two weeks later. There was a lot of unlocks that happened in that room. Just having folks do it, seeing their peers do it and that was I think very, very helpful. From the helping your executive leaders from across NEI understand the value of this really makes it easy for them to support their staff, make it important, help to make sure that the message was clear. We want you to experiment, we want you to try. Failure is okay. The Goal is to learn. There's a lot of misses early on. It does take time to learn how important it is to have that conversation with the AI and get the value you need. The part where I think I would have been a little bit more forward leaning with our executive team early on was we knew we wanted to stand up this Jedi program, our champion program. But making it clear that this is going to require a time commitment from your staff. We definitely want them to you to be very clear that you are authorizing them to carve out time and that there are some expectations if you want to be on this program. We do need to be collecting that feedback. We do need to see you reaching out to your peers, you know, really making it clear because. Because they have a job to do and you're asking them to do additional on top of that. So kind of getting that management level cover there I think may have helped with acceleration a little bit.
- [22:03]
Kerry Weston
So I think what I heard you say was at the beginning, if we had laid out more expectations as to what performance success looks like and what you're saying yes to, rather than just taking anyone with a pulse that said yes, please take me, that would have maybe. Yeah, I understand that completely.
- [22:17]
Bill Gross
Sure.
- [22:18]
Kerry Weston
Not quite an application process, but we're not taking everybody. If you say yes, here's the three or four things we're expecting from you. Please consider that before jumping in. Because it's important that we also have inbound and not just outbound activities. I think for sure. Yeah, that's a great lesson. I appreciate you sharing that because that will fast forward some of the pain or some of the the delays, if you will. For someone trying to do this internally, knowing that that was an issue for you, did you start at the top for a reason? Did you recognize the value of having leadership buy in before we went forward? Was that an on purpose goal for you?
- [22:56]
Bill Gross
It was because I think, you know, I told you I sort of had this aha moment after she came into my office because I had not been using it. And when I realized, and it doesn't take long to do this, when you realize the potential impact that this could have across an entire enterprise, you realize that the core to that change is a culture change. It's not small addition to the way we do work. It's going to touch everything about how we do work. Work processes will change, agents will get introduced. A lot will need to to happen. Even for us to do successful externally. Facing content creation requires a lot of work between different groups that may not normally interface and they all have to sort of understand the role AI is going to play in that. Creating the brand, tone, voice for any eye. So anyone who's creating content is going to be producing content that meets our kind of, our standard of expectation. So understanding that culture change is going to be at the heart of it. You have to start at the top. You can't, you know, it's very hard to change culture from the bott and given how excited Maria was about AI and she characterized it to me as she felt it was her obligation to her employees to prepare them for the AI future that we are all going to live in. And she really feels very, very passionate about it. So if she's that motivated about it, she's going to be my ally and I'm going to leverage that for sure.
- [24:30]
Kerry Weston
Incredibly smart. Yeah, incredibly smart. Great observation and great tactic. And I think knowing, knowing the level of standards details that you have inside the organization, I'd be remiss not to ask you this question. How do you trust what it's doing? Knowing that the information you are putting into it and getting out of it really does have to be accurate and precise? How do you trust what's happening is real?
- [25:00]
Bill Gross
That's a great question. And you know, six, seven months ago, ChatGPT was in a very different place. It has really matured in terms of citations for almost everything it produces and it has gotten much, much more high fidelity in that regard. And I think as because the model developers understand how important this issue is. The other thing is this is a core component of our AI policy is anything that's produced by AI has to be verified by a human. And because we work in such a risk averse field, that's not something that's sort of hard to sell for the group. They'll fact check everything that anybody produces because we're just so reticent to want to introduce an error. So it's normal core business for us, but we have found that with providing the right context, prompting the right way, doing the right checks and balances, the hurdle for validation is not that high. It's definitely, you know, the amount of time savings you get versus the work that you're going to put into validating, you're still saving a ton of time.
- [26:10]
Kerry Weston
Yeah, I think what you're just, you just laid out a perfect case study for what I call the busy middle was you've got your ideation planning phase, you've got the busy middle and then you've got the refinement and you know, the finishing touch. And it sounds like you're just compressing the busy middle to the point where you can put maybe even less time into the refining, but even you put just the same amount of time into refining and editing phase, you've saved yourself a bunch of time in that busy middle, right?
- [26:36]
Bill Gross
Absolutely. For sure.
- [26:38]
Kerry Weston
Yeah. That's fantastic, Bill. That, that's great. I think there's tremendous value in the lessons that you've shared, the things that you've done, the things you wish you would have done. If somebody wanted to reach out to you, if they said, you know, I've been trying to tackle this too, is there a way that they could reach out and connect to you? Because I think there's going to be a couple people that'll be like, I really, I really need to do this inside our organization.
- [26:58]
Bill Gross
Sure. My email is whiskey Romeo Golf, so wrg ei4nuclear energy institute.org email me. I'll be happy to share anything I can.
- [27:08]
Kerry Weston
That's fantastic. Thanks for reaching out. I appreciate it and it really means a lot to me that you took the time to say, hey, listen, I think I can answer some questions and provide some experience and that's what it's all about. So, Bill, thanks for joining me. I appreciate the conversation.
- [27:20]
Bill Gross
Thanks so much, Kerry. Give up the good work.
- [27:29]
T Mobile Advertiser
Marketing is hard, but I'll tell you a little secret. It doesn't have to be. Let me point something out. You're listening to a podcast right now and it's great. You love the host. You seek it out and download it. You listen to it while driving, working out, cooking, even going to the bathroom. Podcasts are a pretty close companion. And this is a podcast ad. Did I get your attention? You can reach great listeners like yourself with podcast advertising from Libsyn Ads. Choose from hundreds of top podcasts offering host endorsements or run a pre produced ad like this one across thousands of shows. To reach your target audience in their favorite podcasts with Libsyn ads, go to Libsynads.com that's L I B S Y N ads.com today.