Get Ready With Me: Generative AI Webinar Prep, Part 1 of 5

Written by

AI, Artificial Intelligence, Generative AI

Get Ready With Me: Generative AI Webinar Prep, Part 1 of 5

In today’s episode, you’ll join me as we dive into the world of hospitality and explore how generative AI can revolutionize the industry. You’ll discover the power of language models like Google’s Gemini to analyze vast amounts of data, uncovering key challenges and potential solutions. We’ll uncover surprising insights from industry podcasts and see how AI can be applied to address labor shortages, shifting consumer trends, and more. Tune in to gain a competitive edge in the ever-evolving world of hospitality!

Get Ready With Me: Generative AI Webinar Prep, Part 1 of 5

Watch this video on YouTube.

Can’t see anything? Watch it on YouTube here.

Listen to the audio here:

Download the MP3 audio here.

Machine-Generated Transcript

What follows is an AI-generated transcript. The transcript may contain errors and is not a substitute for watching the video.

Christopher Penn: This week, we’re going to do a multi-part series on getting ready for a webinar using generative AI tools, a bunch of different data sources, and things like that. I figured we’d do kind of like a “get ready with me” video series — except that instead of putting on makeup and trying on clothes, we’re going to get data, process it, fry it on, and convert a previous webinar into the webinar that we’re going to be doing this week.

Like I said, this is going to be a multi-part series. We’re going to talk about using generative AI tools to quickly adapt a presentation and make sure that it’s relevant for the audience. So this week, Trust Insights is putting on “Generative AI for Hospitality”. Hospitality in this case refers to hotel and hotel-like businesses and the kinds of things that they would need to be doing to attract and retain customers.

So broadly speaking, when I adapt a presentation, I take the existing shell of a presentation and I change the parts that are industry-specific. Let me go through here and show you very briefly what’s in the average presentation.

You have housekeeping. You have the major issues facing that industry, some explanation about how generative AI works, and some basic examples of prompting. Then we get into the use cases: the major use cases — generation, extraction, summarization, rewriting, classification, and question answering.

This is the portion of the presentation where I try to have as many industry-relevant examples as possible, going through and identifying what those are and demoing them. After that, you have “when should you use AI”, “when should you not”, some of the warnings about AI, some of the challenges that will be specific to that industry about generative AI, and then, of course, preparing your career.

So that’s the general flow of this presentation. It’s really well-structured to be able to pick up one portion of it and switch it out with the other. I also go through and make sure we change out screenshots of things as technologies evolve so that it’s current and not showing stuff that’s wildly out of date.

So this is hospitality. The current deck we’ve got here is a real estate one, which obviously is not at all the same thing. So the first thing we need to do is — in the presentation, I like to start each one talking about the challenges facing an industry. So we’re going to change this to “key challenges facing hospitality”. Get rid of the previous generated images, all of which are generated with Meta AI.

Now we need to figure out what the challenges facing the hospitality industry are. There are some that are perpetual. “Heads in beds” is perpetual. It’s what the hospitality industry does. Companies are always looking for how to increase the number of heads in beds, how to increase the profitability of said heads in beds, and how to reduce expenses for serving those heads in beds without necessarily compromising quality.

However, there are specific issues to this industry that we would want to know. How do we go about doing this? Well, we’re going to need two things. First, we’re going to need a snapshot of the industry itself. Second, we’re going to need an ideal customer profile of the major players in the industry: what challenges they are facing.

How do we get that? Well, it turns out one of the best sources for understanding an industry is that industry’s podcasts. Every industry has got a podcast of some kind right now because podcasting is absurdly crazy. Most podcasts sensibly and correctly have stuff on YouTube.

So for the hospitality industry, this is [a] great show called “No Vacancy”. They are up to like episode 892. It’s a weekly show, and they publish a lot of content about the hotel and hospitality industry.

What we want to do is use the closed captions data from maybe this year’s episodes, feed that to a language model, and ask it to just summarize, aggregate, and synthesize what the major issues are. There are ways to do this individually, but you can also — and probably should — do this programmatically.

What I’ve done is I’ve gone and grabbed just the closed captions files — not the audio or the video — for this year’s episodes. Let’s go ahead and clean this up. Now we need to get this into a format that generative AI is capable of processing. This right here is very messy. These are VTT files, VTT transcript files. VTTs are marked up. You can see inside of one, there’s all this extra stuff. Good news is it’s relatively easy to convert that. So let’s go ahead and convert that in Python.

Okay, so what we’ve got now is we’ve taken our files, we’ve gone through the transcripts and gotten all that VTT stuff out so that they’re normal text files, and then consolidated them down into one single text file that’s called “main data” here. You can take a look here and see how it looks. It looks really okay. You can see there’s markup in these files that denotes what episodes or what.

Okay, so now after all is said and done, we’ve taken our huge folder of VTT files. Each one has been cleaned up to remove the VTT formatting, and then each file now has been consolidated into one big text file. You can see the episode title, the transcript from that episode, and then this file goes on for 434,000 words, which is quite a lot. But this is basically the last eight months of this podcast.

Our next step is we need to boil down these eight months into something that is a little more digestible. So let’s go ahead and bring up Google’s Gemini language model. Gemini is my preferred model for this use case because we need the very large context window. Gemini can hold up to 2 million tokens, which is about 1.5 million words. We saw that our transcript file here is 434,000 words. That is about 600,000 tokens. So that means a tool like ChatGPT can’t handle this size data; it has to be something like Gemini.

The first thing we want to do is understand what’s in this file. Let’s give this a prompt that helps us sort of highlight the issues in the hospitality industry. My prompt here is: “You are a hotel and hospitality management expert. You know hotels, customer care, and running a hotel business. Your first task is to read through these transcripts of issues facing the hospitality industry. Once you’ve read through them, outline the issues in outline format. List this in descending order by priority. An issue has the highest priority if it will substantially increase or decrease revenue for a hotel or hospitality company.”

So this is 549,000 tokens. We’re going to go ahead and let Gemini do its thing. Okay, in less than 90 seconds — less than a minute and a half — we’ve had Gemini go through eight months of this podcast to get an understanding of the hotel and hospitality industry.

So in descending order of priority, our major issues are:

Financial and operational challenges: a profitability squeeze as labor costs and insurance costs go up, continued resistance in ADR — I don’t know what that means, so let’s take — let’s see if we can figure out what ADR means in the context of hotels — average daily rate. Oh, so the price of the hotel. Continued increases in average daily rate may face consumer pushback, leading to revenue stagnation. Yeah, the prices are too high. High-interest rates impacting financing, tight lending environment, deferred capital expenditures, pressure to catch up on deferred maintenance and renovations while facing higher costs. Solutions and opportunity: revenue enhancement, cost optimization, strategic partnership.
Labor challenges: talent acquisition, retention, difficulty finding qualified workers, employee turnover, changing employee expectations. Younger generations prioritize work-life balance, flexibility, and purpose-driven work, requiring adaptation and company culture and benefit offerings. Solutions: enhanced recruitment strategies, improved engagement and retention, flexible scheduling, benefit packages that cater to employees’ needs. Give employees aid to see autonomy in their roles.
Shifting consumer trends: experience economy, the rise of alternative lodgings, a focus on sustainability, technology-driven expectations. Guests expect seamless tech-enabled experiences requiring hotels to integrate technology thoughtfully and efficiently. Blurring of business and leisure travel lines between businesses leads to becoming increasingly blurred, requiring hotels to cater to blended travel needs — that is true.
Solutions and opportunities like legal and regulatory challenges.
Targeted legislation over-reliance on hospitality for societal issues.
Brand oversaturation: major hotel companies launching new brands leads to confusion. Brands struggle to clearly define their unique value proposition.

This is a really good, helpful snapshot of the entire hospitality industry and the things that people who are in this industry might be facing.

Knowing this, we can now start to think about, well, what — if anything — would applications be where we could add generative AI capabilities for it. So for example, labor costs and property taxes and the cost of consumables — yeah, there’s no hotel going to be able to use generative AI to address those issues. Average daily resistance — daily rate resistance, consumer pushback — you might be able to do [something] on the marketing front explaining what a consumer gets for their money. So we’re going to bookmark that as something [of] potential interest. Interest rates — can’t do anything about that. Can’t do anything about [the] lending environment or capex. Labor challenges, difficulty finding qualified workers — so labor shortage, this is a big one where, if you were to go into job boards for how companies are trying to hire hotel workers or look at the recruitment strategies, you might find that the way that they’re marketing working in a hotel could use some improvements. So we’re going to put a pin in that for areas where generative AI could be useful.

Let’s take a look at shifting consumer trends: experience economy, alternative lodging, sustainability. A lot of these are travel trend things. This is a case where generative AI could be enormously helpful by listening to consumers at scale and being able to adapt your marketing and adapt your product, your product offerings, using generative AI.

Legal and regulatory challenges — I mean, there’s a ton of things you can do with legal review and stuff. I’m going to leave that alone because that’s all there is to it. 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.

Get Ready With Me: Generative AI Webinar Prep, Part 1 of 5

Machine-Generated Transcript

Comments

Leave a Reply Cancel reply

More posts

Mind Readings: How to Do a Content Audit With Generative AI, Part 3 of 4

Mind Readings: How to Do a Content Audit With Generative AI, Part 2 of 4

Mind Readings: How to Do a Content Audit With Generative AI, Part 1 of 4

Almost Timely News: 🗞️ AI Can See The World (2024-12-15)

Pin It on Pinterest