Tips
Even if you were able to join us at HITEC in Indianapolis, you probably wouldn’t need ChatGPT’s help to guess what topic everyone was chatting about.
Yep, AI.
What’s interesting, though, is how many of us are still struggling to keep up with both the tools and the lingo as this technology advances. Things are changing quickly, but let us a share a quick overview of some of the common terms you might hear, how they work together, and put all of this into the context of a fictitious resort to (hopefully) help it make more sense.
When someone says “traditional AI” they are typically referring to something that is based on a set of rules similar to a logic tree. You take your input, run it through a series of rules, and your answer comes out the other side. This often requires much more human intervention to train the model. It’s why variations of this form of AI are referred to by many different names including rule-based AI, narrow AI, weak AI, programmatic AI, or deterministic AI.
Basic Rate Suggestions
For example, our fictitious resort’s revenue manager might use an AI model that includes a series of rules based on historical performance data. They might input the date, current rate, occupancy, etc. and the model would run those inputs through the model’s rules and then display a recommended rate.
Predictive AI may have similar inputs and outputs to some forms of traditional AI, but it’s much more hands-free and can create more complex outputs. This is where things like machine learning will come into play. Instead of a human looking at historical data, identifying the patterns, and building the rules, a computer will do this all on its own. This type of AI not only builds more accurate rules, it can also build rules that a human may not have anticipated. Even more, it can dynamically predict what might happen based on the patterns it identifies within the data.
Advanced Rate Suggestions & Context
For example, our revenue manager might feed past occupancy, rate, and revenue date for our destination into a predictive AI model. Instead of a recommendation for a specific data two weeks in the future, it might provide a forecast of what your occupancy would be based on various rates you may choose for specific rooms and nights.
Conversational AI is a form of AI that can simulate normal human conversation. Once again, the model is using vast amounts of language data to both understand a sentence that it receives but also create a sentence in response. This isn’t too much different from what you’re doing right now as you read this article. You’re using a lifetime of experience with sentences, words, and meanings to make sense of what you read and, potentially, form a sentence to tell a peer what you just learned.
Rate Suggestion Chat Bot
For example, our revenue manager might feed an AI model past occupancy, rate, and revenue data for our destination and then ask the AI tool, “Are the rates for Aug 26-29 too high?” The bot might reply, “Yes, your rate of $349/night is likely going to negatively impact occupancy. I’d suggest you lower your rates to $329/night to maximize ADR.”
Finally, generative AI refers to the capability of an artificial intelligence model to produce original content. Trained on extensive datasets spanning text, audio, and visual media, these models respond to input prompts by generating outputs that align with the requested format. The result may be textual, such as an email; visual, such as an image or video; auditory, such as music or speech; or a multimodal blend combining several of these elements. In fact, this paragraph was written by AI!
Marketing Campaign Copy & Images
For example, after you lower your rates to $329/night, our marketing manager might ask a generative AI tool to create the copy and imagery for a post on social media promoting these nights. The response might be a graphic promoting a $329 rate plus a sentence that can be used as a caption on Facebook.
Now, keen readers may be wondering which category applies to their favorite AI tool. In this case, the answer is likely all of them!
ChatGPT is a perfect example of this. If I ask ChatGPT to do basic math, it’s doing work that is very similar to what a traditional AI model would do. If I ask it for a weather forecast, it’s going to do work similar to predictive AI. If I ask it to design a logo, it would be doing generative AI. In each case, it will likely read my prompts and respond to me using conversational AI.
Different types of AI are similar to ingredients in a pantry. Depending on the type of dish an AI tool is trying to offer, it may turn to various combinations of AI to create the output their users need and interact with the user along the way.
We’ve released a few AI-based features or partnerships recently. Which categories do these fit into?
Type: Predictive
When a guest makes a booking, our AI-powered upsell engine analyzes their reservation, compares it to thousands of other, similar reservations, and predicts which products that guest is most likely to want that aren’t in their current itinerary.
https://corp.inntopia.com/automatic-upsell/
Type: Conversational + Generative
Our new integration with Q-Concierge allows guests to call an AI agent to book their vacation. The bulk of the platform is conversational AI as you speak back and forth with the agent, but the agent’s recommendations come from generative AI that takes the data it was trained on and creates sample itineraries that the agent speaks and describes with words.
https://corp.inntopia.com/inntopia-q-concierge-partnership/
Type: Generative
While we can’t share too many details yet, a new tool we’re working on was initially conceptualized using AI. We knew what clients were looking for (partly with the help of an AI assistant) and used that insight to write a prompt for a tool that created a rough POC mockup of the feature we’d been imagining.
Does that make more sense?
Hopefully, but if not? All good. Like we said, there’s a lot going on and it’s moving quickly. And if you have other questions? Just ask, we’re all figuring this out together and we’d be happy to help you navigate this exciting, but intimidating, world of AI.
Start selling more of everything your resort has to offer.
Prefer to talk so someone directly? No problem. Just use our calendar links to schedule a time to chat.
