B (3:28)

It, it definitely has because that little bit of a background helps me think about the right questions. Security 1 yeah, and probably most importantly, you know, I had hammered into my head not just the type of, you know, off the shelf maybe trainings that we all get or have to take if you're in a corporate environment, but really working side by side and writing and ghostwriting for cybersecurity professionals. So really starting to know maybe some of what I need to look at to make sure a solution really fits our needs. Security needs up our ourself as WG Content, but especially our clients and the bulk of our clients are in the healthcare industry, highly regulated sector. So that's always going to be important.

A (4:16)

Okay, that makes a lot of sense. So before we started recording, you had mentioned that you guys at WG Content very early on kind of created a, A, a team to explore Generative AI. When did you guys start that? How early into the release of Gen AI did you guys get started?

B (4:36)

Pretty immediately, you know, a matter of months. As soon as it hit and there was traction and just a lot of conversation and you know, you know how that was Initially it was like this thing and it's interesting but what can it really do aside from tricks? Yeah, yeah, we dug into it because there was, even if it wasn't immediately accurate and as early days, it was a lot of hallucination and bias and all of that. I don't think we can ignore what this is doing here in terms of content generation. This may just disrupt us and also might not introduce a lot of opportunity. So pretty immediately our leadership team said we need to focus on that and we had the idea of let's get this group of people volunteer based at least initially. And we felt that was very important too because people as people are still having some conflicting relationships with AI. Right away we wanted it to be people who were willing to raise their hand to felt like they were comfortable or comfortable with being uncomfortable and willing to say okay, I'll give it a try. I want to learn either because I want to push myself or because I'm just highly curious and maybe a more tech savvy person myself, but at least with it not being mandated, you know, you weren't dealing with those emotions initially. So then we developed business cases. This was also very important to us that we weren't, right from the onset, not chasing a shiny object. Probably something else that I learned coming from a technology company. It's really easy to do that. More of finding something to use a tool for rather than having a purpose first and then seeing what tool might fit a little bit a different way. So we look at business cases right off the bat, but really small incremental. You know, I realized this team, we need kind of babysits to just get comfortable with the concept. So there's business cases in the beginning where just what are our minor pain points and what's right in front of us? You know, summarization, note taking. Yep. Using it side by side with email, you know.

A (6:43)

Yeah.

B (6:43)

That type of thing. And then as there became a little bit more comfort or understanding of what it all meant, our business cases grew.

A (6:52)

You know, it would have been a lot harder to get good results when you guys started than it would be for a company today. The models have obviously cleaned their act up considerably. They're much more powerful, much more user friendly. But that, that's something that when we work with companies, one of the first things we advise like, like who are the people who are going to raise their hands if we look for a. An AI council or you know, center of excellence kind of leader type thing. So for the listeners who are maybe taking the shiny object approach and don't realize it, or their company is, is still in the investigation phase of, of AI with the impact that comes from having that kind of core group leading the charge. What would you suggest to them would be like the easiest, best way to start Some lessons learned about creating that, that kind of, that hunt group, that council.

B (7:44)

Mm. One of the things we talked about, and I don't know if this is quite answering your question right away, we'll come probably get around to it. One of the things that I just like hammered right from the start is just don't feel like you have to know everything. It's continuing. It was definitely changing a lot and it still hasn't stopped. And you can feel very quickly overwhelmed with. I don't know what to look at first and I don't understand. You're not going to know it all. So focus, you know, choose your, say LLM, choose your tool and just stick with that one for a while. You don't have to hop around, find one that does the bulk of what you need to do for us. I was chatgpt, who's also, you know, it's trusted, it's highly funded. I don't think it's going anywhere. Agreed. Let's continue to. I'm just, just stay right there. And then my advice from the start is just have it open, you know, so that it's there. And instead of going to Google, you consider going there and you can start there with just asking questions. And again that then you build your faith in it, you start to understand what it can do, you start going there more and more. And that's not to say you can't Google things, but if it's out of sight, out of mind, you'll never break yourself with those habits.

A (9:05)

Yep, yep. Now that's a great bit of advice. Just keep that tab open on the browser, right? Oh yeah. Oh, let me try the. Oh, maybe Chat gp. Yeah, that's fantastic. And also I really like the idea of look overwhelmed by all the tools. It's one of the number one things that we hear from people. Where do I get started? Your suggestion of picking one. LLM and I agree, Chat. We primarily use Chat GPT definitely gets the job done. It makes it a lot easier for those that that team to sync up and get simpatico. For sure.

B (9:38)

Absolutely.

A (9:39)

So you talked about starting with some small business cases. How did you guys go about identifying those?

B (9:46)

Again, you have to look at your own pain points. So you're not necessarily saying where can we use AI, finding a place to just put the tool. What's challenging for us right now? And oh, could, could I maybe help with that? So for us, one of the things being a content creation company agency, we do work with hundreds of different clients. They each have all of their different brand styles and writing preferences that we need to keep straight, that our team needs to keep straight. And they do. But it's challenging when you're shipping in the course of a day and some of it's nuanced and it matters to the client. So you're having to keep track of, here's the tone of voice that this client prefers. And also serial commas, yes or no spaces before and after EM dashes, yes or no. And they hire us to have that ion consistency. So one of the first things we realized is can we make this easier? Can we make it easier to find the information and can we maybe have a tool that does a little bit better than what at that time Grammarly was doing for us? Yeah, you know, not just check for general grammar, but check for the styles of this client. And maybe you serve as like a knowledge base now. So I can go, and I don't have to say, where did we save that file? Click, click, click PowerPoint. But you can go to that knowledge base and say, do I use a space before and after an EM dasher? You know, and it, that alone was such a small thing that sped things up. But that was a pain point for our team, especially for our editors. That was a pain point. And we realized this could help with that. I mean, we actually, like a half a second before ChatGPT hit the scene, we were in conversations with Salesforce about building a knowledge base so that we had something that was more queryable than just going out to SharePoint, finding all these documents. And so as soon as this hit the stage, we were like, wait a minute, this might do that. And it's no cost, low cost. All we have to do is figure out how to build it and that's where the team comes in.

A (11:59)

So rather than some formal process of identification of particular pilots, you worked within the core group, that council, and you guys call it the AI Pathfinders.

B (12:09)

We did. Okay.

A (12:10)

So within that Pathfinders group, you basically looked around the table and said, pain point. What's yours? What's yours?

B (12:16)

What? Okay, yeah, Writers, what kind of slows you down every day? Outreach, you know, content strategists. And that was that initial. And we continue to do that. It's just that our, our pain points, our vision is a little bit bigger now.

A (12:30)

Yep, absolutely. So what were some of the challenges that you guys like had as you started this group up with either internally, like getting buy in from each other or dealing with the executive team or handling, you know, resistance and a change management spectrum from the teams. What were some of the challenges you guys had with this?

B (12:53)

It's unfortunately, leadership was not a challenge. They led the charge and we have a highly involved and integrated, you know, leader that's champion. Everything she attends, she participates, she herself is involved in. Yeah, she herself is a pathfinder. Very interested in learning. So. So that. That's helpful. Change management. That's why. Another reason why we started this group because yes, there were some, those who didn't raise their hand and they had their reasons. Some of them, maybe they, it wasn't that they were AI opposed and they were still figuring out on their own. They just didn't want the added meaning on the calendar. Some of them were nervous. They're like, this is, what's this gonna do? It's rocking the world. This is gonna Threaten my job, especially in content, but as you. So having it come from their peers and not come from me for having them seeing the excitement of in a team meeting, having a peer say, oh my gosh, look what this did. This was amazing. Having one writer say this was amazing to another writer, like, hey, it's not threatening. This is. This helped me today and I'm a better writer, faster writer, whatever, because. So that definitely helped. But yeah, we had to be patient about that. And then within our own team, it's the emotions that I had to help people get through in the beginning. Even on the Pathfinders, I would say the emotions there were, oh my gosh, this is cheating.

A (14:27)

Yes, it is.

B (14:30)

It is so wrong. But, you know, and I was like, just, yeah, keep going. And then, you know, about a month later, they're like, you know what? It's not changing it. Smart. Yeah. But yeah, it's like a little coaching. I'm like, just, it's okay. You're not doing the wrong thing. You're not doing it. You're doing it as an assist, as an assist to your daily work. It's not doing all the thinking for you. It's not doing all the writing for you. Uh, but those were some emotions in the beginning for sure. Also just reminding them that we, we can table an idea and wait for the tech to continue to evolve. It's like it's not quite working now, but that doesn't mean you don't go return to it because as you mentioned, when AI first hit, it wasn't as, as good at content generation, for example, as it is now. We table ideas, but we circle back and we are comfortable. I, you know, I say we do a lot of custom GPTs right now and I say our goal is 80%. Is it 80% accurate? Is it 80? We're not going for a hundred percent. Someday it'll probably do closer to that. But rightmorrow, if we're seeing 80%, that is success and it can still help us in our day. So let the fox and emotions. We had to and continue to.

A (15:48)

So I got some clarity from what you were saying about this team. So you identified. You didn't just say anybody in the company who's interested. You said somebody from this department. Who's interested. Somebody from this department.

B (15:59)

No, we didn't need to because just kind of naturally we ended up with representation from each of those groups. If somebody hadn't raised a hand from one of those groups, I may have tried to just go and ask and encourage again. I wouldn't have wanted to force it. But if nobody say, if no writer was interested, and I went and I said, come on, guys, this is what it's going to be like. You don't want to. And they still said, no, not for me right now. Then I would have at least made sure one of the existing pathfinders went and had conversations with them, interviewed them, made sure we understood their pain points. But yeah, it just fluidly happened that way. We didn't have to run into that issue.

A (16:41)

Was there any acknowledgement, recognition, compensation, advantages for pathfinders to volunteer for the group?

B (16:51)

We do as a company. As a smaller company, we don't have formal performance plans. I worked in some big organizations. That's where it was much more like that. And like you said, it was kind of things were measured and did you do this? And how many things were above and beyond, and then that leads to a bonus or, you know, it wasn't quite like that. Being a small company. It was more about opportunity and people, you know, getting exposure. Yep. Moving your own career along. We're very, as a company, we are just very invested in personal development. Our CEO even provides a stipend each year just for personal development, as you deem fit. So that's already kind of part of our culture. So it didn't have to come with, if I do this, I'm only going to do this because it's going to be a financial gain. It was more of people want to. Because they want to learn and they want the exposure. And then if opportunity does arise. Sure, that's good. On your internal resume.

A (17:59)

Yeah, no, it's definitely not going to hurt. So with this group, at any point did how do we kind of control people's usage of this so that we're not doing risky stuff? We're not putting the wrong stuff in the models. Did that come up and how did you guys address that?

B (18:20)

Yeah, that was immediate. Okay, again, that's that. Oh, I can't not know what I have learned working with cybersecurity experts in the past.

A (18:28)

True.

B (18:29)

So the very first thing we did and what was important, because there was so much skepticism in the beginning. So the very first thing we had to do is define our policy. What's our internal policy that we can then communicate to our clients? And for us, that policy was our policy is your policy. So we will be a minimal. So if client A is comfortable with us using AI in a more integrated way in the workflow for their content types, we'll do that. Column B says Promise me you're never going to touch it. We won't. And so our policy is your policy and that's continuing to evolve. But in the beginning, that was very important for clients to hear, so we needed to establish that. And then also this is a bit of a soapbox for me. Morning. Love it. Oh, yeah. My, my soapbox is like, I see too many companies because they want to get it right and they. So it takes a long time to define that policy, to establish a governance, you know, committee and outline what are we, what's going to be on our can use, can't use tool list, what's it going to block, all of that type of stuff. All that time you're getting that together, people are still using AI.

A (19:49)

Yes.

B (19:49)

And even once you define it, if you don't educate people more on the why it's still on their phone, it's still on their home computer.

A (20:00)

Great point.

B (20:00)

Still going to be tempted. So to just say you're not allowed to use it and end the conversation there, I think is a big detriment that I see in a lot of companies. So we, right from the beginning, for the whole team, not just the pathfinders, said, okay, here's ChatGPT. This is what it means to not train the model. Everybody go in right now, set up an account together, click here, do this something. And if you explore anything else, these are the things you need to look for first so that you know, at least you feel you've empowered people that if they do start to go rogue, they know it's important.

A (20:36)

It's a tricky situation because, yes, perfect is the enemy of done. And while you're waiting, your team is establishing usage habits without you.

B (20:47)

Exactly.

A (20:48)

And if you wait too long, they're just going to be like, well, I already, like, I got a good thing going over here, I'll do it my way.

B (20:55)

Sure, yeah, it happens.

A (20:58)

It does. I know. It makes perfect sense.

B (21:01)

That's why we have that, that concept of shadow it, and you could call it far the way I know, I mean, that's right. Where there's a will, there's a way, especially if they're being forced to do so much in their day.

A (21:12)

So for those business owners and professionals that are listening to this, tell them what shadow AI is, shadow usage is.

B (21:21)

It is just that it's that despite maybe your. And I don't want to scare people, but that it's just the truth of human nature that it's that they're using it behind the scenes. It's tools that they're using. Despite what you may or may not have said, if you haven't come flat out and said you're not allowed to use this, you are allowed to use that. They'll start to explore and marketing, you know, I'm in marketing. Marketing especially. We are, you know, a rugged roll up your sleeves kind of group it. We're always asked to do more with less. And so that's like, well, what tool can do that? And if I'm in a mired in a company where I have to go through it and jump through a lot of hoops like what can I maybe do here on the side that's not breaking the rules but bending them. Maybe it happens. The ZI concept of shadow AI.

A (22:20)

Yeah. So essentially for those of you that think, oh our people aren't doing it, they're doing it, trust me, it's on the phone just like, just like we said.

B (22:31)

Right. So just go ahead and have that conversation.

A (22:33)

Yeah.

B (22:34)

Them what it means to do it responsibly.

A (22:36)

Yeah.

B (22:37)

And you could still say, I mean that doesn't mean you can't put the hammer down on certain tools. Yeah but rather than. But don't just stop there. It should be continual education.

A (22:48)

So once the policy was established was the next step the introduction of these tools that okay, here's the tool, here's the right way to use it according to us. Let's get you tuned up on how to prompt, how to iterate those prompts and things like. Is that what happened?

B (23:05)

Yeah, exactly. So that's some of. Because we have that pathfinders team, they could do some of that early work that kind of hit. Yes. Trial and error. And then when they developed, you know, okay, this, this solidly works in this scenario or. And I've taken it this step. They were the ones that taught back. But it's a continual. We have monthly team meetings. AI's always a topic because there's always something we've learned that we can continue to teach. So again, smarting small. It's like it might start over here in Pathfinders and then it migrates over to the rest of the team. So prompting is one of those. Once we started to get a better sense of what makes good prompting and the rest of the team was getting more comfortable now with ChatGPT. Okay, now let's go back and now that you've kind of dabbled, it's up there, you've got the tab open all the time, you're using a little bit yourself. You could get a lot More of it. Now let's talk a little bit more of a prompting framework. Let's talk about just how many more things you can do.

A (24:08)

I hope you're enjoying the episode. Did you know that while you're listening to this, competitors are already boosting EBITDA with Generative AI? If you're still stuck in pilot mode, margin and market share are leaking every quarter. Cheapaiofficer.com guides executive teams through the three moves they need to be doing. Identifying the top AI plays that will lift revenue or slash cost within 90 days. De Risk adoption with team training, airtight governance and change management. Best practice and then building and deploying production cradle AI powered tools. No extra headcount, full ROI tracking the results, operators at scale. Generative AI cut process costs 22% last year. Let's talk. Claim your free private 15 minute executive briefing by going to chiefaiofficer.com and walk away with a board ready roadmap slots fill fast so lock yours in now. Enjoy the rest of the episode. So was it more of a like a baby step than it was an immersion for the the rest of the staff that wasn't a pathfinder?

B (25:14)

Yeah, yeah it was, it was very deliberate. You know, like we would say start over here, then push it over, start over here, then keep pushing things over. And now we're, you know, whatever, a year and a half, two years into that now it's becoming a little bit less necessary to work in that fashion. Yeah. And it's culture has been established so the whole team now, the whole company.

A (25:37)

So if we've got somebody listening who's a business owner and they haven't gotten started yet, let me just ask you flat out, has it had an impact on the business, this effort?

B (25:45)

Yes. Yes, yes. I mean we all say yeah, I mean everybody who is really using it will say that. And yeah, that does not mean it has to be a big investment.

A (25:57)

Yeah.

B (25:58)

Mean anything fancy you can do. I mean we're a small company, we are doing things we very self sufficiently.

A (26:05)

So are you guys. Do you give teams accounts to the your team chat GPT teams?

B (26:13)

We do. Not everybody has it or at least it's gradually getting to that standpoint because we're watching, we're a small company and we're watching our spend. But those of us who are using it consistently every day and are building custom GPTs, we do have a team account. And for those who may not be familiar, the difference is or the criteria for us one is how often are you using it? Because if we have the free version. You're going to hit your limits. And so if you're in the course of your day, if you need to use it more often, then it's time for a paid license. Yeah. We also then are building our custom GPTs within a team environment because it's that enterprise create security as a default. So we've trained everybody how to set their data controls. But when it comes to building custom GPTs and things that we're sharing with our clients now, tools, you know, these custom GPDs we're sharing with them, we want to make sure there's no, there's less margin for error.

A (27:25)

You know, you've got more of a, a tech industry background than most of our guests. Do you trust that when they tell you they're not going to use your information to train them up like. Because the biggest, one of the things that we hear because we're, we're working with larger companies than when we started now and the only thing holding them back is they don't understand the risk. So what are you, what do you think? Is I tell them, look, the way I see it, my incentives are aligned with theirs. If it were to come out that somebody's information that was supposedly used in a secure environment and chat GPT popped up in a result for somebody else, it would like chat GPT would be out of the running. They would lose all their, certainly hurt their, their revenue. So like for that reason alone I, I trust them. Like they don't want that happening because they're of, you know, they want to make money. What are your thoughts though?

B (28:25)

I, I agree as, again especially as Chachi PT specifically open AI continues to progress, continues to have investments. But yeah, it was a scary time there when they had everybody on like their kind of security team jumping ship. So with that in mind it's, I mean I personally approach it a lot like social media, you know, just because it says enterprise grade security, you know, be a little bit conscious of what you share. Um, yeah, no, if you don't have to put it in there if it's real, probably don't do it unless you have, you know, a version of ChatGPT on your own server. You know, you're, that you're hosting.

A (29:17)

Yeah.

B (29:18)

So for. We've also. That was another part of an exercise that we did early on along with establishing our policy is I went through an ex, kind of a, an experimental workshop with our leaders to say okay, let's just identify what we would be comfortable if, if something got out, what would we be comfortable. What would be, what would be detrimental to our business and what would be like, okay, so for example, I asked the sales team your pricing structures, your pricing models, if for some reason you shared it and you know, you were used with ChatGPT because you were strategizing and all of this. And if, if, if for some reason it got out, would that be the end of the world? They're like, well, probably all in parody with every other agency. So no, it wouldn't. Patient information 100%. Yeah. They have no margin put it in the model. But we are using, we aren't given that type of information when we're working with our clients. So pretty much anything that we touch daily basis I think is fine and safe. Again, I think a little bit. How much do you want the universe to know about you? It's the same concept as social. How much do you want to type in? You can be conscious of that and that's a little bit of a personal decision.

A (30:47)

And you know, whether you're comfortable or not, you just don't have a choice if you want to stay focused.

B (30:53)

Yeah. Right.

A (30:55)

So for those of you listening, I think you've heard from, from two experts just now on that topic. We trust them. Do they? But does what's going on, what's going on behind the door at OpenAI or Anthropic or anything like that? Who knows? But same thing with Google, same thing with Microsoft. I mean you got to be able to.

B (31:16)

Exactly. I think you made the best point, Chris, where it's like it behooves them.

A (31:21)

Don't have a choice. Yeah.

B (31:23)

The company would follow.

A (31:24)

Yeah. The alignment of incentives. Yeah.

B (31:26)

I, I, it behooves them to really do what they say they're doing.

A (31:30)

Yeah. Yeah. So consider those, dear listener, as you consider the risk. It's everybody else is doing it.

B (31:40)

Right.

A (31:40)

And you know, I look at it, in my case, who cares what I'm doing?

B (31:45)

Right.

A (31:45)

I'm not like now this is different if it was a hospital or if it was a law firm or if it was, you know, a drug manufacturer with, with, you know, proprietary formulations.

B (32:00)

Yeah.

A (32:01)

A content company. Like, at least for your internal stuff, like you're a content company. Like we're an AI, you know, training and transformation company. Like I'm not too worried about it.

B (32:10)

So anyway, no. And yeah. And yeah, for us in the realm of marketing, almost everything we touch becomes public. Yeah. Worst case scenario, a minute sooner. But honestly, I really, I don't. It doesn't keep me up at night.

A (32:26)

So that brings up an interesting question. Do you have a different policy tolerance for different departments in the company?

B (32:34)

Well, we are. Since we're so small, it is the same tolerance across the board. I just say that the type of our biggest concern and certainly we're worried about our own content, our own assets, but our clients, because they trust us and that when we look at that, that risk, you know, that line of risk, what's totally detrimental and what we could recover from, losing client faith and trust in us because we've done something with their information that could shut us down. So for us, that's what we look at. But again, and that's why we tend to say in all of our contracts, you really shouldn't be giving us anything. But that's because it's not required.

A (33:21)

Interesting.

B (33:22)

So.

A (33:26)

And everybody listening. We didn't plan on going this deep into like Pathfinder situation, but I think that that's one thing that's applicable to all companies especially. But if you're already started and you don't have it, get one started like get, get one of these kind of specialized groups that's considering this through the lens of your business regularly, like get that started. If you're a company that hasn't started yet, this is a fantastic framework to get you off the ground in a way to where like somebody's watching the pot boil, right? Like it's not going to boil over because you've got people that are dedicated to being the, the internal thought leaders. On the subject of gen AI in.

B (34:05)

Your business, if you don't, you're go, I mean you're gonna fall further and further behind. It's not going anywhere and people are finding results. It continues to improve. And that's really not rhetoric.

A (34:20)

Yeah. So talk to me about the training that you guys are doing. You mentioned that you have a monthly meeting. Is that a separate meeting where AI comes up or is that a dedicated AI training conversation?

B (34:32)

It's a little bit of everything. We are evolving as the need arises. So as it was more infant steps, you know, those early stages, it could be an agenda point on a team meeting. Right. Like here's the newest thing we're learning, here's the newest thing the Pathfinders discovered, here's a note taker we recommend and here's how to use it. But, but as it's becoming more and more ingrained now in our workflows now it's warranting its own meetings, its own true like lunch and learns. And so we have a roadmap where we can put those things on the calendar. We want to very deliberately train our team on this, this, this and this. We have a plan for that based on the needs that arise within our company and the way we're using it.

A (35:25)

So one of the things that I, I have to remind clients or interested people who want to bring some of their company is that quote unquote, AI. It's not an event. It's not like, oh, like we've got. It is a process.

B (35:38)

Yes, right.

A (35:39)

But it's a process that as you introduce it into this, this department, this role, this job, it creates more time to do more. So there's like this exponential opportunity that opens up. One of the things that I'm always curious about is how much of the knowledge sharing is just organic. Like, oh, hey, I see you doing it that way. Let me show you how I'm doing it. Is there a lot of that happening?

B (36:04)

Oh, absolutely. Like, I'm sure every company now we have a teams channel dedicated to AI where people are constantly dropping articles like, wow, did you see this? Did you see that? Or here's what, oh my gosh, I was able to accomplish with this. This with our custom GPT today. Look at this. Or I heard from this client, they're using it this way. It's actually a pretty good idea, you know, but yes, it. You need that space to curve that ongoing conversation. Absolutely.

A (36:33)

I think it's a great idea. So anybody listening? If you don't, if you've got the. The AI conversation is happening, but you don't have a channel for that where everybody gets to learn the lesson at the same time. I think that would be a fantastic idea. Meetings are great, but things get lost and as fast as AI moves. Like just yesterday, I mean, our chief AI officer is on top of stuff. Somehow I found out about agents chat GPT agents.

B (36:57)

Yes.

A (36:58)

Like two minutes after they started going live. We've got a community of chief AI officers that are kind of like always. And that was so helpful because a, I was able to bring that, you know, bring that to the. To my chief AI officer rather than the other way around, but also like, there's people that won't know about that and won't start testing it for. For weeks just because even though the person you know and the other office may be using it, but if there's not that, that water cooler for AI chatter, I think it's a missed opportunity as these people share stuff. Because the people in your company, they understand your culture and they're all looking at AI through the lens of that culture. So that conversation is extremely calibrated to AI in your business. So I would encourage that for anybody.

B (37:46)

Absolutely. Great advice.

A (37:48)

You've mentioned custom GPTs a lot, and a lot of the stuff that we do, the work with clients, the training, the certifications we do, one of the first things on day one that gets taught for sure, because it's so powerful, I don't know what the usage is of like how there's what, 5, 6, 700 million weekly users of Chad GPT. How many of them have bought or, I'm sorry, created or used one that was built a custom GPT, I don't know. But once you get like, once you get it dialed into that 80% output level you're talking about everywhere, Bing, bing, bing, bing, bing. How was that process for you? Because I know that they weren't like they chat GPT 3.5 didn't get introduced with custom GPTs. What's that process been like for you guys and what's been the results?

B (38:39)

Well, because as you mentioned, you know, we've had that one person, couple people who can at least always be seeing what's, oh my gosh, this just dropped. So it pretty much as soon as that capability dropped, I was like, oh my gosh, that's interesting. And I can translate to how we could use that. So yes, right away, we're our own best testing grounds. We do it first internally and then share it with our clients who their hands may be tied or, you know, they're caught in their day and it's just not within their culture to have space to learn about stuff like this. So we did, we started before we.

A (39:23)

Go on, Brian, I guess for those that may not know, can you explain your definition of what a custom GPT is?

B (39:30)

Good point. It's a feature within ChatGPT and it's really, there isn't much to it, but you can. I call it a, you know, it's a tailored version of ChatGPT, tailored to your specific need. So for us, that can mean that we train it on voice. The voice that how you want it to talk, how you want it to write. Preferences. It could be writing preferences. Preferences like please always, you know, always produce results in bullet point form, things like that. And the goals. What is it I want you. What's your role? How are you going to help me? Think of it as, you know, you just gained an extension of your team. You just gained a personal assignment. Yeah, that's how I look at it. What do you want me to had, you know, custom GPT to do. So when you're looking at that, there aren't a lot of components to it. You give it some instructions, think about is that a long form glorified prompt and you give it some knowledge. What, what do you want it to base on? That's that, that's the part that's tailored. So our very first, you know, GPT that we built for ourself was in bring a knowledge base. So let's put in our process documents, let's put in on brand guides and here's the voice I want you to have. Here's if someone is asking about questions about this, go here, if that go there, you know, it just gave it a little bit of a structure and understanding of what to do with the prompt when someone asks it a question. But then for our team it was first just that, just a place, then ask questions and it's pulling the answer from all these documents that we have in there. But then like, okay, it knows our brand. So now I can use that as a writing assistant and it's also going to write and for me, who's one of the few non writers on the team, I can use it to analyze my writing and make it sound more like one of the writers on the team in a consistent brand voice. So now you have more people on the team able to speak with that same voice, able to communicate consistently and that, that builds your, your brand and reputation as a company. So that was kind of our starting point. Yeah, but it's growing from there. We built, we come brand gurus, we built those same type of things for all of our clients. But also we have a content strategy assistant that knows kind of what we know or what are our personal best practices. Now we can ask it to analyze our strategies, analyze webpage and it's going to do that with a filter of what's important to us campaign testers. You kind of give it our Persona. And so in marketing we talk about Personas. It's think about in the absence of a focus group, here's this fictitious person that you can now pitch things to. It doesn't take the place of people. It would be better to have a focus group, but sometimes the time and the money is not there.

A (42:42)

Yeah.

B (42:43)

So while I can pitch some ideas and say, let's say I'm writing a, I don't know, a blog about, I don't know, about pregnancy, about high risk pregnancy, I can pitch it over here knowing that my intended reader is going to be maybe a 40 year old woman, maybe she's scared. Maybe she's, you know, this or that. Okay with that mindset. Analyze my blog now and tell me did I miss anything? Are there questions you would have liked to know about that I didn't include here? And it's amazing how much it can perform on that level.

A (43:19)

Two things I like about that. One, anybody in the company can now like get. They can produce results from their role as if they were much senior in their role than they are. I like that a lot. I'm curious about that first GPT. Was it intended to be almost like a user guide for the company for employees? Like this is how our company does stuff so that anybody in the company could access that GPT and kind of have it run through the boss's lens kind of.

B (43:50)

Right. Well, not even that. Just. Just the culture of the company. Just find information and maybe right from that voice of, of the brand. That's not necessary. I mean the CEO is a reflection of brand likes to see. True. But you know, it. It's what's your mission as a company. So it's always riding in that, that voice. Coming from that perspective.

A (44:13)

Yeah. So we're got a few more minutes left. But there was another topic that you mentioned early on before we started recording that I think is super cool. It's everywhere now and it's very important and that's this AI version of SEO Search Engine optimization. Everybody used to be able to like keyword stuff on their website and pop up when somebody did a search for, you know, a Thai restaurant near me or something like that. But now fewer. There's certainly been an impact on search volume because people are going to or they're just staying in chat GBT rather than bouncing between Google or Perplexity and then chat GPT. So what has been kind of your exposure to it and what are your thoughts? Is it, is it here? Is it real? Can you game the system like you could with SEO?

B (45:12)

It's here, it's real. The gaming component, you know, to be determined. Everybody's trying to figure out. But I, I don't know that it's going to be as, I don't know, I say simple as it was with SEO. But SEO did not feel simple. It took many, many years for strategists to figure out what makes Google click. So for your audience who may not be mired in the marketing content strategy, what we're talking about is SEO search engine optimization. That's one of the things that our clients come to us at WG Content. So Traditionally you Google something, you put in a keyword or two into your search bar and you get results, you get all of those blue links, all those different articles that you could choose to click on from there. Well, what we do as content strategists say, okay, not only are we going to help you write the page so it reflects you, it's, it's informative, it's well written, but there are some tactics that we can employ that make your information more likely to show up as one of those links. So that, that's SEO. And through a lot of trial and error and some leakage and tips from Google, eventually content strategists have got that pretty down to what do you, if you do this on a page and you refresh it constantly so that Google sees that it's, it's current and yeah, you'll, you have a better chance of showing up high on that list. And the higher you show up, the more likely someone's going to click through your website. So AI is different. So now when you use AI search and the best example, if you haven't yet used an AI tool like ChatGPT, if you haven't yet used it in this capacity, just think of those AI overviews that pop up now in Google. That's that summarized column at the top that gives you an idea of where things are turning. So instead of now, okay, here are some things that mention that keyword. Now it's I can ask a really long question, multi part question and it'll summarize it like okay, well based on all of these different sources link, here's the best concise way to answer your question. And from there then the user can go in and have a conversation. That's the difference. Now I can say, oh, interesting, okay, tell me more about this or okay, let's go a little bit deeper on that company you mentioned or you know, what are their ratings? You can just keep going on and on and it's not a new search, it's a conversation. Yeah, it's not you. Okay, let me put in some more keywords hunt and pack around some of these articles. It's just there, right like that. It feels more, it really does feel like talking to a person. So how do you optimize for that? That's where we're at now because as more and more people prefer that method and they are every day and I traditional Google search will go away. Yeah, if not entirely, it'll be very slim. It's hard to imagine because that is so ingrained as A behavior in our.

A (48:37)

Mind 20 something years.

B (48:38)

Yeah, hard to imagine. But all it takes is putting ChatGPT in that and your browser, tabbing it and using it in this way and you never want to go back. Why would I want to hunt and peck for answers when I can have a summary there? And initially when it first came out, the bite was. Yeah, well, how do I know it's accurate? Well, now there are citations. So if I doubt, I can click and I see what the source was and I can double check. Sure. But anyway, so how can you possibly optimize now for that search when it's pulling from a lot of different sources? How can I be one of those sources it pulls from and one of those that it actually cites? That's where we're at now and that's what everyone's trying to figure out. But it's not going to be as easy a game. It's not just a set algorithm like Google had or do. This, this, it appears and it ranks. Now it's really because AI is paying much more attention to like, well, what's the nuance of your question? What's the intent for what we're finding is instead you still need to do all of those things that we've done all along for traditional Google. But now, in addition, can we write with thinking about like, okay, this answer instead of me building a big long story form and you have to kind of read the whole thing to get the gist, can I make sure it's summarized? The gist, the answer, the immediate answer to a question, a thought is summarized. So then AI comes along and it's like, oh, that's packaged neatly. Let me pick it up and over it over here. Things like that are what we're starting to play with. It just subtly changes the structure of your page a bit. And that's just one example.

A (50:27)

Well, I'd like to stay in touch on that topic because I know you guys are producing a lot of content and you're not just producing it in a vacuum. You guys are paying attention to clicks and how many people are, you know, forwarding and all like all the, all the metrics that come with the impact of content. And I would imagine that SEO prior to today was probably something that you guys considered as well.

B (50:48)

Oh yeah, that. I mean, we were, we were experts in that. Yeah. And we still are. It's just that SEO is now AEO Answer Engine Option optimization as well. That on top of it means how do I frame my content so that it's more likely to show up in those people also ask this. Yeah, how do I make sure? Like, okay, I've considered what people will also ask. And then there's now geo, Generative Engine Optimization. And that's like, oh, that's where I was talking about how can I make sure there are kind of some like concise summarizations or concisely and concisely answered questions that could be lifted up out of the framework of an entire blog, say or out of an entire web page. But really SEO, aeo, geo, I just say it's all going to become, it's just AI optimization. It's just all of those things combined now.

A (51:50)

So, Diane, I don't know if this would be any of interest to you or really anybody listening, but we have a free community at chief AI Officer and we've had a few of our chief AI officers that have done kind of a deep dive and done some tests. So if you're interested, I'll send you a link. It's free to join.

B (52:07)

Yeah.

A (52:08)

And you can catch. And the people who taught those classes, they are, they're not goofing around like they were very serious. They explain it in, in non technical terms. But might, there might be some nuggets in there for you, for you guys.

B (52:21)

I, I would love that.

A (52:23)

Okay.

B (52:24)

As, as the theme of the day is, none of us can know it all and we all, it's, it's a.

A (52:30)

There'S a lot of noise, but there's.

B (52:31)

A village mentality and that's what I loved about this, if anything. Yeah. Be afraid to ask questions. Just nobody's got, nobody knows, nobody does. But we're all willing to share. Yeah, it's kind of nice that in that realm it's not so competitive. It's like, hey, we're all just trying to figure this out. Yeah.

A (52:50)

Hey. So Diane, if individuals want to find out more about kind of what you guys are doing with AI or just in general at WG Content, where should they go to kind of keep abreast instantly?

B (53:01)

Follow me on LinkedIn. I try to post pretty regularly about some of the ahas and things we have, we discover along the way. But you can also go to wgcontent.com and we have a service line offering called WG Content Catalyst. And that's where more of our AI activity falls. We have things like adoption, AI adoption workshops, probably a lot of similar things that you're talking about where. And you know, if you just need to catch your team up, you know, we can customize some of that experience some of that education around you specifically, certainly when it comes to content. But even in general, like, how do I get a Pathfinders team started on my own?

A (53:45)

I like that idea a lot. So we're going to have those links in the show notes here. And as always, like, the best thing that anybody could do after listening to this or any episode is open up Chad, GPT or your LLM of choice and start beating up the keyboard. Start asking the questions.

B (54:02)

Yeah. Just. Yeah. Awesome.

A (54:04)

Diane, thank you so much for being a guest on the show and the information. It didn't go the direction that we thought it might, but I'm really pleased with the conversation that we had about.

B (54:13)

I love it.

A (54:13)

Pathfinder Concepts.

B (54:14)

Yeah. Thank you for giving me a chance to, you know, 100 talk about this stuff. I love it. It's.

A (54:19)

And I'm gonna make sure that the team sends you over a link. And for those of you listening, we'll add a link to join our Chief AI Officer community. Again, it's free. Fantastic content that we produce in there and that we distribute widely and freely. So. All right, everybody, thank you so much for listening on the episode and we'll catch you on the next one. Thanks for tuning in to Using AI at Work. Don't forget to subscribe for more conversations about how to use AI at work.

B (54:47)

Work.

A (54:47)

And a special thank you to our sponsor, Chief AI Officer for empowering businesses with AI education and training. Visit their website for a free AI readiness assessment and AI strategy guide to help you get started using AI at work. That's www.chiefaiofficer.com. so thanks to our producer, Evan Desaunier for making this episode possible. Follow us on Twitter at the handle using using AI at work and visit www.usingaiatwork.com for free resources to help you harness AI in your role.