I'm working through a project with ChatGPT and our content. Thought I would share the progress and some findings that might be relevant.
Today in health it data governance and its impact on your AI results. That's what we're going to talk about today and we're going to use. My experience to, , discuss what it's gonna look like. I think what it's going to look like in health systems. My name is bill Russell. I'm a former CIO for a 16 hospital system and creator of this week health, a set of channels dedicated to keeping health it staff, current and engaged.
We want to thank our show sponsors who are investing in developing the next generation of health leaders, short tests and artists. I checked them out this week. Health. Dot com slash today. As you know, we've partnered with Alex's lemonade stand for our five-year anniversary. And we have a goal of raising $50,000. We're just short of 40,000. We'd love to have you be a part of it. And we thank you. And those of you who have already been a part of it, you can participate, go to our website, just underneath the banner.
You're going to see the logo for the lemonade. Stand, go ahead and click on that to give today. We believe in the generosity of our community. And we thank you in advance. All right. So a couple of things. One is we are starting our AI journey at this week. Health. , we've actually been utilizing chat GPT for some simple things that we're doing, like writing our show notes.
So we will do a recording. With a guest and we'll have that, , we'll have that transcript created by AI. We'll drop it into a thing called descript and it creates the, a transcript and it'll put, , markers for when I'm speaking. And when the. , when the guest is speaking. And we use chatty. P G chat GBT to go through that. See the content of that. , identify two quotes from our guests.
And create the show notes. Four. The, , for the show. And so that will be on our website. That's also part of the email that goes out. That's, you know, that's basics, that's blocking and tackling. , we are in the process of automating that somewhat through Zapier and some other things, but I'm going to leave that for another.
Another day. , because this isn't about automation, it's really about our use of AI. So we just had a 2 29 event as part of that 2 29 event. We decided to get our whole team together since. And it's most of us we're going to be in one place. We float flew in the remainder of the team and we met on Wednesday. And one of the things I did is I had each of the members of my team do research on a different AI project.
So give you an example, our graphic design person. Who is phenomenal. , Jeremy fallen sits in the background, does all our graphics designed for us and you probably never hear his name. Probably may not even know that he's there, but, , he gave a presentation on the various tools that are available.
, graphic tools that are available that are using AI right now. And he went in depth in what Adobe is doing with regard to AI. And it's really interesting and really fascinating. , the person who he gave to go through chatty PT. T for us, came back with an interesting proposal and it is a product called custom GPT. Now, a lot of you are doing this. You're doing this through your relationship with Microsoft and you're signing up for Azure.
GPT or whatever they're calling it. But essentially your own sandbox of your information so that you can query it. Using the chat GPT interface. We're doing the same thing. It's called custom GPT. We can give it an XML site map of our website. It can go out and peruse all the episodes that we've ever done.
And then we essentially embed that onto our website, which I've already done. If you go to our homepage, , I'm not a stickler for getting things perfect before they go out there on the site. , if you go to our homepage about a quarter of the way down on the homepage. You're going to see custom GPT. It's going to say, ask any questions.
And I'll be honest with you. A lot of the questions you ask right now are going to say. I need more information. Because it crawled those pages, it looked at those pages. And the information it does have, it's really exceptional. Like you can ask it, what did Craig Richard Bell say in his last episode? And it will tell you what Craig Richard Bell CIO for Intermountain said.
And his last episode, because it went through those pages and those pages have the show notes on them. And it has the, , , the transcript as well. In some cases it's linked to the YouTube video, but it doesn't automatically go and listen and download the YouTube video transcript. So for that, you have to give it another set of links that it'll go over to the actual.
Page for the YouTube, , , video and grab that. Transcript as well. Very powerful tool. It uses GPT four, not. GPT 3.5. , so I liked that about it. I liked how easy it was to set up. I essentially gave it an XML site map. And it went and browsed, , all of our pages and it will continue to update those pages as they are updated. Just like a Google crawl of your website. Would.
Here's what we learned in the process. And I think this is relevant to health systems. It is. Data governance matters. I don't care how good the AI tool is. Your data is going to feed that tool. And it can only answer questions from the data that it has. So one of the interesting things that I'd forgotten about, we just ha I just, hadn't been keeping up on this. We have a small team. There's a lot of things that, you know, we have a list of things we would love to get to if we have time and we never seem to have time.
Well, one of those things is we didn't start. Doing transcripts until probably year. , gosh, you're three, maybe. Of the show. And so we're in year five now. So we have two years. Of good solid shows with transcripts, and then we have three years without any transcripts. So my expectation was I was going to feed all these pages in.
And I was going to get this really powerful tool to ask it any questions. Tell me all the things that BJ Moore, CIO of Providence. Has said on the show, give me an outline of that, that kind of stuff. And it didn't do that. And I was like, well, why wouldn't it do that? BJ has been on the show three times.
And it really only has information from one of those times because there's only one transcript on our website and that's an example of data governance. What information are you going to feed these tools? You get Azure GPT. What information are you going to feed that tool? That it can provide value back to your clinicians, back to your health system. This could be a really interesting front end to a lot of data. Queries could be a lot of really interesting front end.
Into. Anyway, it could be a fascinating front end. Conversational front end into a lot of your data. But it has to be able to consume that data. First of all, it has to be available. Second of all, it has to be in a form that can be consumed now. The beautiful thing about these tools is it consumes all sorts of stuff. It consumes PDFs, consumes videos.
Consumes edge. So it is a really powerful tool from that perspective. But it's going to matter the data you feed it. It's going to determine the answers that it provides. And so subsequently the value of the information that it provides. And so I wanted to share a little bit of my learning on this. As we move forward, we had a 2 29 event last weekend in Charlotte.
, 13 CEO's in the room. , all of them are doing something. With Azure GPT chat, G P T. , we had a great conversation on AI governance. And I, you know, I think that a lot of health systems are looking at this. I thought I'd share with you some of my insights as I head down this path. And learn. And my goal is that that homepage custom GPT would become more and more valuable. My vision for this is essentially every interview we've done every today show we've done every Newsday show we've done in custom GPT, and then you can query it with simple language that you would normally use.
And say, you know, has bill ever interviewed somebody on the topic of, , oh, our efficiency. And it would provide back some information. The cool thing about custom GPT is it says here, you know, the source of the information, you can actually click on the source and then see, oh, it's these three episodes and then click on those episodes and listen to those episodes or read the transcript.
So that's the vision for it. And so we define the vision. We defined the objective. And how we're going to deliver it. And now we are feeding it information and we have to go back and clean up the data. As is usually the case in some of these things. When you, when you get that new technology and you're ready to go after it.
, invariably, what happens is you realize the mistakes you made along the way that are going to make the tool less effective than you would like, and you have to go back and clean that stuff up. And so that's what we're going to do. And, , I would love for you to go onto this journey with us and let us know if this tool becomes more and more valuable to you. I would love it to be a repository.
, of what CEOs are saying in the industry. A place where you go to right off the shoot. And say, , you know, and, and ask a questions to get very specific. , details of what. People are saying. About the topics that we are in the middle of E right now. So the toppest we live every day. All right. That's all for today. If you know, I'll get back to the dues. I'll get to the news tomorrow. I just thought this would be interesting to share plus.
I spent an awful lot of time today on this project. So I thought it would be relevant to share. But that's all for today. If you know someone that might benefit from our channel and you know what to do, send them a note, let them know that you're listening to the show and you get value out of it. And you'd love to talk to them about it. They can subscribe on our website or wherever they listen to podcasts.
We want to thank our channel sponsors who are invested in our mission to develop the next generation of health leaders. Short tests 📍 and artists, I check them out at this week. Health. Dot com slash today. Thanks for listening. That's all for now.