Get Ready With Me: Generative AI Webinar Prep, Part 2 of 5

In today's episode, this is part two of my get ready with me series on how I prepare for a webinar using generative AI. Let's continue. In a lot of ways, the ways the companies and corporations tend to address uh legal and regulatory stuff is trying to circumvent regulations and do so in a way that is beneficial to the corporation, but at the expense of the general public. And this is true across industries. It's not unique to the hospitality industry.

It's everyone. But in the interest of using AI ethically, uh, we would like to discourage people from using it to uh to manipulate the law any more than they already are. People will anyway, but I don't have to tell them that in a webinar. So we'll skip past that. Brand oversaturation.

Now, this is a this is a pretty easy one where dealing with your brand and your marketing. How do you do brand positioning? How do you do stuff like that? So this is um this is gonna be uh some good useful things. So we've got our major issues.

Let's go back to our presentation now. We're gonna set up um some of the use cases we talked about. So let's talk about dealing with recruiting issues, brand uh what do we want to call this brand differentiation? Umployee engagement. Well, see the thing is this is a generative AI for hospitality with a focus on marketing.

But we I think employee engagement might be useful there to to deal talk about that. Okay. Recruiting issues, what else we want to do? Experience economy and technological innovation. So those are five major challenges, and I think we can build some decent use cases for all five of those.

So let's go ahead now. Uh the next thing we're gonna need is we're gonna need some uh imagery to populate this. We'll use Meta's AI, because uh the Meta AI generator generates some pretty decent flip heart like stuff. And very critically, for people who are concerned about the ethics of using image generation models. Meta's models, for good or ill, are trained on data from Meta and publicly available data, but data from Meta, especially things like Instagram and Facebook.

Meta can prove that people have signed the terms of service agreeing to allow Meta to train on the data that we give it. And so from an ethics and law perspective, Meta's models are probably more compliant and are not infringing on anyone's intellectual property, as opposed to say, like stable diffusion or open AI, where they did not get permission to scrape images off the web. Meta got permission, perhaps not knowingly, uh, and perhaps to some as someone to say coercively by saying you can't use Instagram if you don't give us permission to train on your data. But at least there is a paper trail that says, yes, we had permission to use all these images you uploaded uh to Facebook over the last 20 years to train our AI. So the image quality you'll get you'll see some fun stuff, but we'll we'll go ahead and build some some of the art for this.

So our five areas are recruiting issues, brand differentiation, employee engagement, uh, experience economy, and technological innovation. So we'll just go ahead and make some images here. Okay, so now we've got our very nice um little image chiclets of uh the different major issues that we want to talk about in the webinar. Uh recruiting issues, brand differentiation, employee engagement, the experience economy, and technological innovation. Uh, we're going to remove this slide on sales hacks because we don't need it for this talk.

Uh that was for real estate and sales. Our slow screenshots of our major services still look pretty good, so we can go skip past that. Uh, let's see, Watson X. This is a screenshot is out of date. So it is time to update that.

Okay, this is now up to date. We can see that it has the current models that are available in Watson X. Something like that. You would think it's not a big deal, but it is one of those things where if you're making a talk to an audience, and potentially in that audience is someone who has experience in the field, in this case, generative AI. Uh, it's a good idea to make sure that there are little nods to the level of expertise and that you should have for a talk.

So having screenshots that are you know two years old in generative AI would be a pretty obvious indicator that you haven't updated the talk, and if the material isn't fresh, maybe the ideas aren't either. Keeping even just things like screenshots relatively fresh is a great idea for subtly communicating that, yeah, this this thing is up to date. Let's continue through our regular stuff on uh the way LLMs work uh is still pretty good. Although I think I might want to do a new dump truck uh picture at some point, but not today. Uh, that's a photo from the State Library of Victoria.

That's generated let's see prompt engineering so I like to do the word cloud activation here um for basically explaining how models synthesize uh outputs and then we get into our three most basic screenshots which talk about prompt length uh these get updated for every single talk because it's just a basic good idea and we'll talk about things like hospitality and hotel marketing so let's go over to Gemini here. So in this next section what we do is we do three different prompts. We do a very short prompt in Gemini what are the best practices for hotel and hospitality marketing in 2024 and get a very generic bland answer. We double the length of the prompt by being more specific best practices for independent hotel chains in the New England region of the USA, especially the Boston metro area get some more stuff and we go hog wild and create a long prompt uh that sort of details what you're doing as as if you were the hotel owner like hey here's the stuff we're trying out and of course you get a much more specific result and this is all part of the section on here's how to uh here's how to to use these tools these AI models to get better results so we generate these each time uh specific to the industry uh we go past our race framework and our pair framework there's no need to update those those are up to date prompt libraries up to date and that brings us to use cases which is the next section of this and this is where uh I like to have and I'm going to take a screenshot of our initial five challenges because we're gonna need to reference these challenges in our use case development for these slides. And it's important to note that for this process of making these uh presentations, if we want them to be credible, if they want them to land with our audience, our audience has to see themselves in the presentation.

We can't just do a generic presentation. We have to use the knowledge that we've accrued to simulate what it would be like to be that customer. And in doing so, we'd be able to uh to create examples that resonate with them. And so we need to have those those use cases handy uh as we as we move ahead. So, our first example for generation, let's reference our list.

Our list is things like recruiting issues, differentiation, employee engagement, experience economy, technological innovation. What would be a good generation example? And perhaps something on the employee engagement side of things. So let's create a let's go through a prompt of how would you generate even just some ideas for this. I'm going to start off with the pair framework, which is uh available from uh Trust Insights.

If we go to trustinsights.ai/slash pair, you can download a copy of the pair framework there. I'm gonna put in my canned prompt, and my prompt says uh let's explore best practices for employee engagement in the hospitality industry, uh, recognition and appreciation, development and growth opportunities, open and transparent communication, empowering employees to autonomy and decision making, work life balance and well-being, common mistakes uh less experienced practitioners make, underestimate the value of development, failing to communicate effectively, micromanaging a stifling initiative, neglecting work life balance, things generally believed to be true but are actually false. High salaries equals high engagement. It's solely HR's ability, uh responsibility, etc. Things generally believed to be false but are actually true, tangible ROI, small gestures of appreciation have a big impact, and so on and so forth.

Lead by example, celebrate success, empower employees to be problem solvers, never stop learning. So, an example for hospitality would be to generate a checklist of useful practices that would be appropriate for a hotel chain. So let's actually go ahead and ask for that. So, my prompt here is from these best practices uh and our conversation so far. That is it for part two of my get ready with me series.

Stay tuned in the next one for part three. If you enjoyed this video, please hit the like button, subscribe to my channel if you haven't already, and if you want to know when new videos are available, hit the bell button to be notified as soon as new content is live.