I LOVE AI With Noelle: Get Ahead of GenAI Risk and Create your own Value Proposition GPT

Instagram, the Amazon Live, like, now's your chance. Oh, see, I saw some people like, ah, click. But let me go ahead and change this really quick. I'm just gonna take 5 minutes and then I'll give it back to y'all, I promise. But then thank you for reminding me, Renee, because I would have been very sad to have to wait another week.

Me too.

That would have been so sad. So thank you so much. So I get— many of you know I get boxes of stuff and then I open them. I also still have my Amazon show. I'm not opening this today. But how many of you could sit with a brand new product and not open it? I'm waiting because I want to open it with y'all. So we're not doing it yet. We're gonna do this next week. But it haunts me. So this box has been sitting here for like a week. And now I want to open it with you because it's a present. And I love opening presents with people. So I'm gonna spotlight myself so that we could see. So welcome. Hi, everybody. If you're joining us on LinkedIn, we are here to do a bit of an unboxing, but But also, we're going to see some of our AI leadership leaders shine today. So it's an opportunity to hear from some of our leaders. I'm super excited about it. But before that happens, I was reminded that I have some presents. So I can't wait to open it. I have to— like, it's taped really well.

So that's good.

All right, here we go.

Oh my gosh. Let's do it. All right, here we go.

Okay.

Oh my gosh, it's already beautiful. There's a box. What is happening now? I have not opened this, so I'm hoping that— what is— what? It's like moving around in here. I have a feeling I know what it is.

Wait.

Oh, oh, I have to untie it. My bad, my bad. This is so fancy, so fancy. All right, I'm almost there, almost there. He's so cute.

Okay, hold on.

Come on. Okay, this is, um, Christmas lights. It's a mindfulness activity. You just have to breathe and like untie it. There, I did. I did. Oh my gosh, it's so soft.

Of course.

So Renee has a book, Shadow— what is it called? Shadow Hunter?

Shadow—

Shadow Stalker. Shadow Stalker, and it's a wolf. What? It's gonna sit right here with all my, my tigers. We have to create like a baby wolf analogy for the AI security, and then we have to build an AI security course. You know, Renee, we're doing that 2026.

It's happening.

Yeah, such a good idea. Oh my gosh, there's so much I want to talk to you guys about.

Okay.

Um, all right, cool. This is very pretty. Thank you. This is like my kids are like, I love the box. I'm like, no, no, no, there's something inside of it.

Oh, What?

Thank you.

All right, there's— oh my gosh, you're like me. You like wrap and then like everything is wrapped. I love that because I enjoy the gift.

Smells delicious.

What? Oh my gosh, I wish I had— I just bought a bag that is this color. I'm opening presents, honey. Oh my God, don't show it.

Don't.

That's awesome. I love it. It's my color. Is that your color too?

It's really cool.

Purple is amazing. Yeah, I love it. Oh my gosh. Oh, this is like made for me. Oh my gosh, I love it.

Tell me about those pants. Are they super comfortable or something?

Yeah.

Well, purple.

Yes, they're purple.

Oh, okay.

I got you.

Thank you.

They match my book.. But more importantly, you know, I don't know, I'm just very tactical. So when I, I always like stickers, like I will draw things multiple colors. I just, you know, it's the creative part of me. I was a scrapbooker. Were any of you scrapbookers back in the day? And this is my, probably my favorite one. It's a hard copy because I bought the Kindle version on unlimited. I did read it multiple times. You get extra credit.

Yes, you do. And there's a couple of stickers in there too. Wear your cape to work.

I love it. Oh my gosh, this is so cool. What? You have such great handwriting. Holy mama, I'm gonna have to take a lesson from you. That doesn't even look like real. I'm gonna grab that like in a screenshot later. Thank you so much. All right, but wait. There's more.

That one's for your children.

This is amazing. I didn't know you— I don't have this book. So this is How to Talk to Anyone with Ease. We might need to buy this for people in our own community because I feel like all of us struggle with it. I mean, I like talking to people, but it was a hard-earned skill over like a decade. So, oh, I love it. This is so cool. Oh my gosh, thank you. Okay, so last one.

Thank you.

I love it all. So we're excited. And so I'm gonna try and stay focused here.

This is—

and there's just— yeah, and then the other, the FedEx box.

Yes.

Alright, let's do it. And I will open your cards too. What? Oh my gosh, they're so cool! That's so cool! Thank you so much, you're awesome! Oh my gosh, and I love stationery. So I just— I mean, some of you know this because you've gotten letters from me, but I just like hand— I mean, in the world of tech, right, that we are in, in the world of AI automating everything, there is nothing like A handwritten letter, a handwritten note. It is, you know, I mean, I still have notes that people have written me and I hang them up in my office because it means something. It's so important.

And my friend, my friend handmade those.

Oh my gosh, they're so cool! They're like textured. Wow. Okay, I see what they did there. Wow, that's pretty fancy. This is amazing.

Who's—

you have to tell me your friends that I can actually Anyway, it's so amazing. Thank you very much. So some of you might get a, a thank you letter, a thank you note for me, and it'll come on this. So really, really cool. Thank you so much.

I love it so much.

Thank you so much, Renee. You made me so happy. I absolutely— and the reason I like to show it to you all is just, number one, it is so good to give stuff away. I try to give stuff away. I sent you all who were part of the Vibathon 2 weeks ago. I sent— hopefully you got it from Sherilyn. I'm just waiting on 2 more people to fill it out, but I have a— you're all getting hoodies who built something on the Vibathon. If you weren't there, you know you missed it, but you're getting I Love AI hoodies, and I just love giving stuff away. It's just fun. You get something in the mail. It's amazing. And so thank you, Renee, for letting me be on the receiving side because I love giving, but it's fun to receive too.

Sorry, Noelle. Good afternoon, everyone. So I'm asking about the hoodie because I received a link But I need to look for— but I need access. I did not get to see the access till now. I'm not sure.

Okay, I will go in and give you access. Uh, it should have been a public link, but I will give you access to make sure. Maybe, um, the email that I registered was not the same email as you ended up. Yeah, so yeah, I'll give you access, no problem. Well, good, you're one of my two, so great. Thank you for telling me that.

Okay, thank you.

But yeah, so if once that is done, uh, we got your sizes and your address, and you will see it show up in, in Right now we are in a beautiful window where you go get to get it before Christmas. Um, and so I'm excited to do that for you all. So thank you for being with me during the Vibathon. It was a lot of fun. I'm really excited, and I can't wait for you all to, uh, to sport your new hoodie like I am. It's fun. I love it. So thank you all. Thank you for everything. Um, but now is really the reason that I showed up today was not to talk to you and of course share with you all the goodies that I got today, but also to give the floor to two amazing AI leaders. And I'm really gonna let you all kind of take it from here, and Skye's gonna transition one speaker to the other. I'm gonna go off camera, and I'm just like, it's not about me, it's about you all. But don't forget to tell us like your story. You know, sometimes we jump right into the thing, and I'm sure, I mean, Joe, you talk all the time, but I just wanna let you know this audience, that, that is here wants to know not just what you do but who you are. And I'm so excited to learn more about both of you. So I think we're going to start. Did you guys have a person? Did one of you think you were going first? Because if not, I'm just gonna pick, you know, I'm gonna go. Oh good, Jill will be the first one.

Okay, great.

Thank you, Carolyn.

Thank you.

Um, all right, Jill, so you are up. Hold on, let me, um, adjust, uh, really quickly. I just need to un— oh my gosh, my children are screaming. I'm glad Zoom can't— you can't hear nothing.

Can't hear anything.

Beautiful.

I love it.

Okay, awesome. So let's see here. Jill, where are you? Where is my face?

Hold on.

Sorry, it is not— oh, there we go. Found it. Okay, great. So now, yes, Jill, you are up. I'm going to spotlight you if that's okay. And then that way we'll also grab this snippet of your presentation, give it to you, of course, so you could share it. More importantly, we'll post it in the community just in case anyone misses it, or more importantly, you want to hear it again. So thank you so much for being here, Jill.

Yeah, thank you so much. Oh my goodness. You all know, I think, the world of you in this community. So it's really my gift to be here. And just thank you so much. This was a fun way to wrap up the week. But yeah, well, to Noelle's point, I'll just give you a little backstory about myself and then why it is I'm talking about this particular topic. But I started my career in tech, really managing tech teams who were building websites. Got to know kind of from beginning to end how to build a website. Yeah, it was primarily, yeah, for web and that kind of thing. Sorry, I'm getting flustered. This is such an honor for me. But what I did most recently is I was working in a digital agency as a research director and really got into responsible AI in particular. I was really just thinking so much about the users and the potential, like, ways that things could go wrong if we're not mindful and intentional about how we build. And so I was fortunately put in a position where I could partner with our AI engineers, our data scientists, and the teams who were doing delivery, both on the agency side where I worked and also with our clients, some of whom were Fortune 500s and so forth. So what I did with them is I designed an anticipatory design session. And we would do this before any generative AI project that we would take on with a client. We— it was a non-negotiable. It's like, if you're gonna engage with us and you're gonna do some proofs of concept, you're gonna go through this exercise to really get grounded on what it is that we're trying to build. But then also think about, okay, what are the risks that we might encounter here? And constructively, How can we mitigate those going forward in a really tactical way? So I was super inspired by Noelle's recent, like, Responsible AI certificate. Many of you, I think, participated in that. And I think I even told Noelle, I was like listening to her while I was folding laundry going, yeah, that's right, that's right. Because exactly what she is saying is what is, is so important for us to bear in mind to be good stewards of AI. And this approach really gives you hopefully a way to make that very tactical. And you can use bits and parts of it if you don't want to use the whole thing. I would do this with clients. It could be anywhere from like a 3-hour exercise to a 3-day exercise. So you can go pretty deep. And if you want support or guidance on how to do it, I'm happy to, but I'm just gonna give you the lay of the land for this lightning talk. And another kind of fun side note is I became aware of Noelle when I presented this as a workshop at GovAI like last year. And it's amazing anyone came to my workshop because there's a lady over there with all this attention, and I'm like, what's happening in that room? But we had fun in our workshop, I would say. So let me share my screen and then kind of walk you through. Oops, I'm gonna send a share request to turn that on. And I'm gonna hide myself view, I think. Nope, here we go. Now I can share. Perfect, perfect. So Here we are. Here is the rundown of the framework. So, before you would engage with clients for this exercise, I think it's important to do a little bit of table setting first because the people aspect of it is just so important. So the way that I frame this is as a trustworthy AI workshop. I just believe that we have to earn trust. And so I think the word trustworthy really kind of signals that in our choices and how we build. So the other thing that I have found the hard way is that if you try to have a conversation about responsibility or trustworthiness in the abstract, it can fall really flat because all of a sudden people start sort of ideating about, oh, what's the, you know, you know, worst case scenario? We go down these rabbit holes and they don't feel productive. They just leave people feeling a little like, meh, like, what do I do with this knowledge? But if you ground it in a use case, if you say, you know, this is what we're going to build, this is who we're going to build it for, this is a little bit of the tech stack that we're thinking of, you can have a really constructive conversation. I also think, you know, having psychological safety as part of this exercise. Now, if you have a short amount of time, it's hard to just kind of bring that with you, but you can lead by example or just simply say, you know, we're going to be asking some, you know, maybe difficult questions here, and that's okay. You can model that with how you engage with folks as well, but it is really important because sometimes people don't like to talk about risk. They might be, you know, thought of as kind of like naysayers or something like that. And we, we need those people. We need those voices, particularly early on. So that's all to the good. Another thing to be familiar with is the NIST AI Risk Management Framework. And I am happy to share a link to that. But I have a link specifically to the trustworthy characteristics. That are outlined in that framework. That's kind of like— I'm not gonna say checklist, but it is kind of a guide to help you work through the variety of risks that you might encounter so that you're not anchoring on like one or two. You're kind of covering your bases. All right, so that's the setup. What happens when you get into the workshop? There's a few different activities and conversations that I like to have. The first is that I like to think about you know, really digging into the use case. What's the whole point of this? Because if you're building an AI system that hasn't really reconciled the user needs that you have with the business needs, it's going to be hard to build something that you can, you know, kind of accurately assess is working as it's supposed to. And therefore, you know, you've mitigated some of that risk. And so it's just an opportunity to make sure everyone's on the same page. You know, who are our users and our major subgroups? What is it we're hoping to accomplish? How will we know when we get there? And how do you reconcile any tensions like, you know, maybe we wanna build a system that is highly personalized, but we would have to collect tons of data to do that. And maybe our users, this is maybe a sensitive context, maybe we, you know, they would feel uncomfortable divulging that amount of information. So we need to kind of find the right balance that's really honed to the need. So use case, use case, use case, problem definition, super important. Alright, and then this is a little bit of secret sauce, I think. And if you are in the design world, you will look at this empathy map, and I know it's small, it's cool. I've got links to it that I'll share. But what I want you to take away from it is really more about the questions that we ask at this stage. But if you're in the design world, you might look at this and think, oh, that's pretty old school. I mean, like empathy mapping, people have been doing that for a long time. But I'm going to tell you, I have been working in these premortems with deeply technical teams along with product managers and some of those folks. And I will say, 9 times out of 10, they have not really stepped back to think deeply about the users who are going to be engaging with the system. And there is so much nuance that comes out of this conversation alone. I warn people that just by filling out the pieces in this map, which I'll explain really briefly, you are not doing the work of empathizing with a user, but you are definitely getting your headspace in the right way to do that. And so maybe you'll think of questions that you're gonna have to dig a little deeper into to really understand their needs, but this just helps you kind of get an entree into it. So the kinds of things that you would ask in this session, and I just have people put sticky notes up all over the place. What do they need to do? What do your users need, or what do they see, like, in their environment? Like, if you put yourself in their shoes, so to speak, what are they doing, saying, thinking, hearing? And it's not really about putting things in the right category, but it's about, like, putting yourself just in their full context. And then you start thinking about their inner world. What are they trying to achieve? What are they afraid might happen? What are the things that they're trying to gain out of this system? And again, I just, I know it's a simple thing, but I guarantee you do this exercise and I think you'll be pretty amazed what people come up with. And it's just aha moments for folks. The next thing I want to talk about too is the tech stack. So I said that knowing the technologies you anticipate using is useful at the outset of this. It doesn't mean you're not going to change as you, you know, sort of prototype, go through solutioning. You may very well make adjustments.. But if you know directionally that you're going to be using some natural language processing or generative AI, that's really helpful because each of those different technologies carries with it its own little risk profile. So we know with generative AI, we're thinking about things like hallucinations, maybe overreliance, you know, unpredictable outputs, things like that. So that technology has its own little bundled set of considerations. So we want to think about the method. And we also want to bear in mind, like Noelle talked about this a ton in our Responsible AI course, data sources are another— I mean, that's like the bread and butter of these AI systems. But really stop to think about where are we getting the data from? What was it originally curated for? Are there any kind of biases or missing data in the datasets? What else might our, you know, proposed solution be connecting with? These systems don't, you know, exist in a in a little bubble. They're usually connected to other systems in the ecosystem. So we wanna bear that in mind. And then finally, I think about like the user interface. You might not be at this level of detail yet because this is still usually early on in the development process. But you might also wanna think about, you know, when, and I'll show you in a minute when we get to the trustworthy characteristics, if you're thinking about transparency, well, the user interface and how people understand what they're engaging with on the front end really has a lot to do with whether or not they experience harms and risks. All right, so the next piece is that NIST Trustworthy AI set of characteristics. Now, um, specifically, again, these are trustworthy characteristics. The entire framework encompasses way more than that. It gets into governance and monitoring and all of that. But I anchor on these characteristics because they're fairly easy to understand and kind of, uh, walk through for people, and they're extremely well-defined in the NIST framework. So I just recommend getting super familiar with that, you know, you're probably gonna have some to spend some time helping others get up to speed with what they mean because there are just nuances with, um, what does transparency mean versus explainability versus interpretability. Those are kind of nuanced terms no one uses particularly consistently, but just make sure you give them an opportunity beforehand or during the session to at least kind of know directionally what they are. Um, NIST has a whole rationale for how they lay things out, and I'm not gonna go into the definitions, uh, in detail, But I will say validity and reliability takes up so much space here because it's foundational for every other, you know, trustworthy piece. People will not trust a system if it's not valid and, you know, does what it says it's going to do. Likewise, accountability, transparency, those really carry over into all of the other characteristics. So those are some nuances you can talk about in addition to the fact that you are often going to have to make trade-offs depending on the use case and your risk tolerance. So If you maximize for transparency, for example, you might then be, you know, thinking, or you might be in effect creating a system that maybe is less secure. If it's completely transparent how it works, someone could potentially game it. So, you know, that's a conversation that you'll have as you do your ideation. Which leaves us to the last stage in the cycle, or in the framework rather. And I'm packing a lot into this one square, but I wanted to talk about grouping, triaging, and planning. Because what will happen at this stage is, you will have gotten really clear on the problems you're solving, the use cases, the users who will be affected, what their perspective is. You will then understand the technologies that you're thinking about and the risks. And in this, previous phase, you will be mapping these trustworthy characteristics to everything that you know about what it is you're gonna build, everything we just talked about. So what you do is you, you know, invite folks to think of as many risks as they can across each of the categories. I usually try to tell people, you know, don't over-index on maybe safety or privacy. Like, challenge yourself to put at least one sticky note in each of these categories and ask yourself the question, what could go wrong in terms of safety, knowing what we know about all of those other things? What could go wrong about privacy, validity, and so on? So what happens is you end up with sticky notes that have themes, and you'll, you know, usually be able to see within each of those characteristics, okay, we really have found a lot of concerns about reliability. Like maybe hallucinations in this use case are just really problematic, or we have concerns about that. It's a good time to talk about why that's concerning. You know, what are we worried about happening, you know, to our users if things were to go astray? And just getting clear on risk tolerances and things like that. And then what you would do with that information is put it into risk categories. So what's the likelihood that this could happen? And then what would the impact be on the users? And I will tell you, there's not one way to triage these risks, but you can think about, you know, how the, how much variability, you're seeing across all of these. Is there a ton of different risks across many different dimensions, or are you anchoring on just one or two? And so maybe you wanna prioritize that one. Think about the particular vulnerabilities of your users, your legal and regulatory environment, even your values really come into play here. Like maybe there's only one risk in maybe validity, but it's really important that you get it right. So you won't know going in, but this will help expose what's important to you. And then what we finally do is we plan. Because as Noelle Russell would say, you know, we are ultimately builders and doers. We are not just talkers. And it's very unsatisfying to sit with this information feel overwhelmed and then wonder, okay, what am I supposed to do with this? So what we would do with our groups, and again, this is a cross-functional group. I don't know if I mentioned that, but as a prerequisite, we have our technologists, our product people, our leaders, as many different functional roles as we can to brainstorm. How can we make this right? Okay, we're worried about validity in this case for these reasons. What could we do from a technical standpoint, from a user interface standpoint, from a policy standpoint to prevent the worst from occurring? How will we know early on if something's going astray? And so this is your opportunity to do a bit of solutioning and start lightly mapping that to your technical architecture, which is what we've done. So ultimately, by going through all this, and it is a lightning talk, so I know I'm going fast. But what we were able to deliver to clients was sort of all of this guidance around clarity about their users and use cases and risk profile. That really exposed what their risk tolerances were and then be able to deliver to them. And here's what you do with that. Here is your plan for going forward with a safe, reliable, responsible AI technology. And I found it worked so well that, gosh, we did it, you know, very often. But what I heard from groups that made me feel like I was really making a positive difference in this space was that they would say, you know what, I never thought of this stuff before. Like, none of this had occurred to me. And I'm thinking, oh my gosh, I mean, these are, Fortune 500 company teams in high-risk industries like financial services, and it never occurred to them that these things could, could be problems because they're so solution-oriented. That's great. But, you know, we, we have other accountabilities too. So that is my whirlwind tour of this framework. As I said, I delivered it for GovAI. I'm happy to meet with anyone who wants to go into it in detail. There's a lot of nuance about how to structure it, but I hope that you find this a good way to implement a lot of the responsibility principles that we've talked about in our classes. So that's it.

Amazing. Thank you, Jill. So I personally learned a lot. Like, I heard of trustworthy AI, but seeing your presentation, it's like, oh my God, it's set for tangible small bites for us to really ingest it that way. So in the visuals there.

So thank you.

Next we have Miss LaTarsha, and let— correct me if I'm wrong, you're going to do how to create the GPT, right?

Yes, yes, yes, yes. So I will start sharing now.

Okay, one moment, let me spotlight you.

Okay. Okay, all right. So I just really want to say I really love hanging out with all of you guys. You guys are my peeps, my friends. And what Jill talked about is amazingly tied into something I want to talk about today, which is about creating this experience that's human-led for others. So I'm going to talk about this common problem that we have. It's the— we're all in here, we're all studying, we're all becoming and are becoming AI practitioners, but we have this mystery that surrounds us is how do I put what I want to do, what I do, what I say, what I stand for into words? And so that led me to talk about what I'm going to talk about today. So that's why we're here, finding the right way to describe who you are, who you serve, and what you do. But, but first, let's talk about where it all started. Yes, you got that right. These are my buddies, the Flintstones. And so I started early being inspired by, like, the gadgets and humanity meeting technology. And that's really where we are today.

There was—

I was fascinated by human innovation, and this gave me my start just for today about where I'm headed, about human-first and heart-centered AI. Now, I don't know, some in the audience may know about this gadget. So when I was about 8, again, my fascination with technology and how humans interact with it started really early. My story surrounds me and my 80-year-old grandmother. See, the tape recorder started, and I'm gonna bring something into view, and if you were there in that era, there was this need for a pencil or a pen because tapes would get stuck. And as she tried to use it, I saw the struggle even then of what happens when we start going down a path of technology and not ensuring that everyone can use that technology, which is why I love hanging out with all of you and everything we're learning in AI Leadership and I Love AI, because the human core of what I'm gonna talk about today, it's about human-first AI experience, even as we are designing apps for ourselves or apps for other people. So this is the core of what I want to talk about today, which is the confusion. What do I say? How do I stand out? Who am I? See, we're all evolving as technology is evolving, and having the right words to say that connect with other humans is where I like to think about and spend my time with, which is why I'm going to talk to you about how to create your own Value Proposition GPT. I'm going to walk you through the steps I took. I'm going to share with you the file, and I'm going to go over even to ChatGPT to show you the buttons I push to get it all started and to work. Then I'll talk to you about how you can get just the thing that I'm talking about today for yourself if you want it, or you can build it. So why is it so important? Because as we are all growing and evolving with technology, we have to know how to talk to our potential clients or people that we work with about AI and what we do and, and who we're there to serve. And oftentimes when we don't get right, or we struggle. We see eyes glaze over if we get too technical, and we find ourselves just spinning our wheels on how to get people to understand what we're doing. And over time, it begins to erode our own confidence and make us feel a little bit shaky about, well, how do I even talk about what I do? So this is a way that you can approach this. I'm going to show you how to build a GPT-4 for this purpose, to actually help you structure your thinking and your thoughts and turn your vague understanding of what you should say into clarity for yourself. It's about removing buzzwords, and it's really about connecting human to human about how you solve problems and what you're passionate about and what you do. So we're gonna first talk about the 5 Ps of framework because that is what guides us into whether we're building a GPT for ourselves, a GPT for others, any other AI system. We all have to go back to these framework ideas. And the 5 Ps are understand the purpose. So when you see my GPT document, You'll first see that I'm going to define what it will do, and I'm going to define who it's there to serve. I'm also going to give it any kind of knowledge or expertise. So if you're, let's say you're doing a, creating a PromptGPT for yourself, this is where you'll say, here's what I love, here's what I hate, here's the people I think I want to serve. I love the thing about ChatGPT and GPTs in general is because you can type and you can say, here's what I like, here's what I love, and here's what you need to know about me, and have it ask you questions to help you fill in the blank. Then also a part of prompt engineering is making sure that we are clear with our AI implementation or AI tool that we know, tell it exactly what it's supposed to do and exactly what it's not supposed to do. For instance, if you're using MyGPT for value proposition, then I'm going to make sure Chat knows not to answer questions like, how long is the Nile River? Or who's today's president? That's not the purpose. And so I want to talk about, even if you're designing something for yourself, to keep your— to keep these 5 Ps in mind because you're going to want ChatGPT to be the most effective for what the goal you have in front of you. So let's talk about the purpose. My purpose for the building of GPT read something like this: This GPT helps entrepreneurs identify and articulate their unique value proposition by asking targeted questions. Challenging generic statements and guiding them to specific authentic differentiators that resonate with their target market. And so I was very clear with this GPT about its purpose to make sure that it gave non-generic responses. And I had to say, this is what's most important to me as people are using my GPT. The next was me spending time understanding my audience. So I did some research. I, I gave it research that I had went out and had it perform research for me and say, here's what I'm looking at and here's the people I'm looking to serve. Can you tell me what their pain points are? Can you tell me what their knowledge level is and what their goals may be as they are trying to work with my actual GPT. This is important, again, whether or not you're building for yourself or you're building for other people. The next is this knowledge base. You can place in there, like, how you want your GPT to sound. For instance, I don't see— you may have a target market that you want to build your GPT or this particular GPT for that are a part of the scientific community. And so you would have to say, these, this is the framework by which you need to understand my target, not my, my target market, who my customers are and what they're needing and what they're needing from this tool. Now there's some examples you give it to talk about, here's what I do need you to answer and here's what I don't want you to answer. Answer. So that was a general overview of what, how to use ChatGPT and what I start using it to do. And what I'm going to do now is let me escape from here and I'm going to stop sharing because it looks like I, and I'm going to share again, um, because I was sharing that. So I'm going to now go transfer over to ChatGPT and show you how all of this works once you're in chat.

And Latasha, is it something where you want us to follow along and go into ChatGPT too to test it out, or is it just a demo that you're gonna do?

It's a demo that I'm doing, plus I can also share this part for others if they're interested in, and then I have a GPT that I've developed that if you wanted to go get it, you can go get it for like a dollar and test it out yourself if you want to. Okay. But I wanted to show the file that I had set up that kind of talks about what the purpose of the GPT is. I'll make this— sure, this is a little bit bigger— what the purpose of the GPT is and also the context. Here's where I place like the undesirable conversations. Here's where I place, um, I'm an expert. Like if it asks, you know, can you help me with my taxes, is to answer, I'm an expert in helping you craft your value proposition statement. I can't help you with that. So I have given it instructions on what it should and should not do. I've given it instructions to give an introduction. Hi, I'm here to help you create a custom value proposition. And then I make sure that I tell the AI engine what to do. Greet the user, support them, give them answers. Then I tell it what questions to start with. So for me, I'm saying step 1, start with you. What comes naturally to you? What do people thank you for? So what I can do is I can share this doc in the— I can share now in the chat with everybody. And you can now have this and follow along.

Now, you know, we all love free stuff.

So I'll share it with everybody after I finish sharing my screen because it's looking like I can't share. So this is the actual GPT that I built. And I'll tell you what I do after I go through and after I ask it to ask these very specific questions. So here, if you have other questions for yourself or for somebody else that's using this tool, you can add those questions here. But if you notice, I tried to keep it short and easy for my users to interact with. And then at the end, I allow the user to say, what should the message actually feel like? Choose these various words. Like for some, some may want a GPT that sounds relatable or calm or encouraging. So I let the user pick even just the tone of how they want it to sound. And then I ask it to give an output. So I tell it what formula I want it to use as it's trying to get this, build, make this GPT work. So from there, what I do is I go over to ChatGPT. When you get this, you can download it as a text document. That's important. You can go over to ChatGPT and you name it. So let me, before we over here, let me back up and show you where I go. I go here to Explore. So underneath the GPT in Chat, you can go to Explore, and then you click Create. And from Create, you can actually name your GPT. So here I'll just say Demo GPT, uh, Value Proposition. And then you can add a description. Now I'll tell you this about this very specific field. It's a short description, but what it really means is 300 characters or fewer. And so it will error out after you've gone over 300 characters, but I do like the fact that it gives you that many characters to describe for yourself, to remind yourself what it's for, and to describe for other people. Now, this is my favorite part, is because we spent that time building the file, the instructions that I give it is really, "Read your instructions. Because this file gives it the instructions for how to actually execute this GPT. This is beautiful. I can say, get started. These are conversation starters. So I can say, get started. Ask me question 1. And whatever I want, and whatever makes sense for my users, to use if I want to say menu, show me the menu. And then I'll upload the file here. I'll go and grab it from my downloads. There it is. And here we are. And now ChatGPT used to generate a photo, but now you can go and grab, let's say I have a photo of something over here. You can go grab a photo and it'll place this photo or icon in your GPT. So if I say create, it will create and notice what it gives me as an option. It says I can only, I can make sure that it's tied down to only me. I can give it to anyone with the link., or I can place it in the GPT store. In the GPT store, I do share my GPTs with other people, and I actually have a GPT store that I started on Gumroad with a variety of GPTs that I created. So I always use anyone with a link, and I say save, and from there it gives you a link. Now it's important that you can copy this link. I copy the link to keep the link. And anytime you lose the link, you go back and navigate and you can say share from up here. As you can see, you can say share. And now I can now view the actual GPT and use it. So it's getting started. The goal is for it to start asking the questions. So notice how it's doing exactly what I told it to do and it's asking the questions. And now you can get started using it and begin building out your own I help statement. Are there any questions? Yeah, when you were creating it towards the end, when you said you showed the menu, like getting started, and then it said the menu thing, you said get started, ask questions.

I didn't understand the menu part, the next slide.

Okay, let me, let me share this with you. For some GPTs, if a user is using my GPTs, I'll go back over here and I'll just type menu. Let's say you're in the middle of this GPT and you've gotten lost in what it can do. You can type, the user can type menu and it now will give you, the user, an understanding of what this actually is doing. That's awesome.

Thank you.

I appreciate you describing that. Thank you. So I like the fact that now in the instructions, I've probably said when the user types menu, give them this information. So this actually helps you to understand especially if you create a lot of GPTs, what this GPT does, and it asks your— it allows your user to go back and understand what this GPT actually does. Any other questions?

Well done.

All right.

Paste that link.

All right.

So now I'm going to— let me find the link. And let me know if you can actually— I was trying to paste it in here. Okay. Let me know if you can actually see the link. Okay. I'm trying to paste the link in here. So what I can do is paste it in the community if it doesn't paste. Oh, there it is. It pasted.

Yep.

I got it. But it says request access.

Okay, let me go and change the sharing. I'll figure out how to get this to— I'm copying again. All right, here's a new link. And there you go. Now you should be able to use it.

Yes, ma'am.

I have a question. Thank you for your presentation. So can someone, without any ChatGPT installed in the phone can get access to your link and, you know, and navigate the GPT you created? Oh, can it? So once they get the link, you're asking whether they can now get your file? Yeah, no, if they can access the GPT even if it's not installed in the phone. Oh, yeah. So I'm going to go ahead. Let me I'm going to share this GPT, this demo GPT with you now. And what you'll see is once you give someone access, so I'm going to share my screen one more time so you'll know how to give people access. So let me see if I can share.

Okay.

Let me share. So I'm sharing again. And I go back to my GPTs.

Yes.

And I go back to Explore and I say my GPTs. Now you'll see that I have built several GPTs here and I have a whole other account where I have like a host of GPTs that I offer in my store. And what I can do is I can just click here and edit and click share. And remember it gave me this option When we were creating it, it says anyone with link. If I copy that link and I now share this link with everyone in the chat, which I'm doing right now. Now you can go and if you click the link, it should populate in your very own ChatGPT instance. Okay, okay, okay. So the thing I like about this, if as you share it, the link actually helps others navigate to chat, it opens up chat and says, okay, you want to start now. Once you get the link, you want to save the link because sometimes it's a little bit difficult to find that link again. The other thing I like about this is as you want to tinker with your GPT more. Anytime you tinker with your GPT, for instance, let's say I don't like this photo and I want to change to, let's say, this photo. That's a, another, that's a better photo. Let's say that's a better photo and I like it. I can just say update. It updates my GPT. And then anyone that has my link has my updates automatically. All right.

Any other—

what I also do, like when I get the GPTs, I pin them to my sidebar. To LaTasha's point where sometimes you don't remember what the person's GPT name was, so I always pin it to my sidebar. If I access it too, so then it keeps it in your top GPT section there on your left-hand panel. But this was absolutely amazing, and this is actually one of the first times where we had two people present on the iHeart Community Call. So thank you, Jill, and thank you, LaTarsha, for your presentations. I learned something, and it's exciting to learn from other people, hearing your expertise and insights around how to create a GPT, the responsible AI implementation side. I like how Latasha, you pointed out the 5 Ps that we've been learning for, you know, building bots and agents and stuff, and now you apply that to the GPT itself. So this is very, very unique, and you gave us some freebies. So who does not love free stuff? But I know that we're in the, approaching 5:00 PM, so I just want to say thank you to everyone staying on, participating. If you all see the chat, it's lots and lots of love, and I appreciate this community. And we hope to see you next week on Pat Meaney's adoption call on next Thursday, and then we'll be meeting again next Friday for this call. With that, you all have an amazing weekend. Thank you.

Take care, everybody.

Thank you.

Bye. Bye-bye.

Hey, Mama Sky, hope you have a great weekend with your little people and the holidays coming.

Thank you, ma'am. Am I going to see you this evening?

What's this evening? What am I missing?

The, um, AI Solution Architecture course.

Guess what? I already am certified.

Well, you know she's teaching something new again.

No, I know. There's only so much Zooming. It's kind of like crack. I mean, seriously. If I develop heart condition, I'm going to be submitting a community bill to get me back to, to put my bypasses, pay for my bypasses. So yeah, no, I'm trying, I try to walk around when I'm on a gazillion Zooms, but yeah, no, I've got a, I've got some proposals out there and it's Friday night. And so, as you know, I want to put guilt in the minds of the people who have not responded to me.

The pressure, adding the pressure.

Yeah, no, it's, you know, you develop a good relationship and then they say, give me a proposal. And then suddenly.

They're ghosting you.

Yeah, then they have to, you know, they, they either put their money where their mouth is in terms of the interest, or beware. I'm that old that I know this game so well. So anyhow, but really great people, and one way or another do something at some point. But anyhow, um, your hair looks great, by the way.

Thank you.

Because now we know the way to Noelle's heart. I'm just going to send her a bunch of cardboard boxes with prompts and then secret websites to turn her on.

All right, guys.

Oh, we still have people on. Bye-bye.

Yep.

Did anybody else have anything? All right, well, I'm ending the call.