I use ChatGPT daily to help me with everything from which restaurants to try when visiting a new city to which artists to see at a music festival, both based on things I’ve previously told ChatGPT I like. Although I’ve used ChatGPT to plan trips, I’ve been hesitant to give it access to my loyalty accounts and travel bookings.
However, Gondola AI — the company that provides data for many of our monthly hotel points valuations — recently launched an MCP (Model Context Protocol) integration that works with ChatGPT, Claude Desktop and more. In short, MCP is an open standard that allows AI applications to connect to external tools and data sources, such as Gondola.
I’d already given Gondola access to scan, parse and track travel-related messages in my email accounts, so I decided to try Gondola’s MCP integration in ChatGPT to see whether it’s a tool I should use going forward.
What can the Gondola MCP do?
Gondola offers MCP tools for travelers, developers and travel advisers. In this article, I’ll focus on the Gondola MCP tools designed for and available to travelers.
The most compelling part of the Gondola MCP is that, depending on what you’ve shared or connected, it can know your travel preferences (such as home airport, preferred hotel chains and preferred airlines), loyalty account details (including status and points balance), upcoming trips and the details of your previous trips.
Gondola says its MCP integration lets travelers use ChatGPT (or another supported AI client) to do the following:
- Search real-time paid and award availability across Marriott Bonvoy, Hilton Honors, World of Hyatt, IHG One Rewards, Wyndham Rewards and more
- Compare cash rates to points rates for hotel stays, including calculating cents-per-point values
- Predict whether the current rate at a hotel is a good deal based on historical price trends
- Get a checkout link for hotel rates you find (although Gondola facilitates the booking, the reservation is made directly with the hotel, so you can still earn hotel points and elite-night credit and receive applicable elite benefits)
- Create rate alerts for specific hotels and dates, and get notified when the price drops
- Obtain information about hotels, such as amenities, reviews and location data
- Search real-time paid flight fares by route and dates
- Search real-time rental car rates across Hertz, Avis, Enterprise, Alamo and more
- Check the credit card rental car coverage by card name
Even if you aren’t particularly interested in most of the Gondola MCP capabilities, you may find it adds some useful personalization to the results if you’re already using a supported AI client like ChatGPT to help you plan or book travel.
Related: Using AI to write complaint letters? Here’s why it doesn’t work
Reward your inbox with the TPG Daily newsletter
Join over 700,000 readers for breaking news, in-depth guides and exclusive deals from TPG’s experts
Setting up the Gondola MCP for ChatGPT
If you are a ChatGPT user, setting up Gondola’s MCP integration is extremely easy. Just head to this page on Gondola’s website to learn more, and then click this link to install the ChatGPT plugin. You’ll see the following information, which you should read and understand, before clicking the Authorize button.
Connecting Gondola to ChatGPT does give the AI client a new way to request information from your Gondola account, so you should review the permissions before authorizing it. Gondola says access is granted via OAuth and that you can revoke it at any time from the Connected Agents page in your Gondola account.
If you’re using ChatGPT through a company workspace, your administrator may restrict which plugins you can install or use. This was the case when I tried to install the plugin on a company-provided ChatGPT account. But I had no issues setting it up on my personal ChatGPT account.
Related: Can Hilton’s new AI planning tool help you find the perfect hotel? We put it to the test
My experience using the Gondola MCP in ChatGPT
Gondola suggests several prompts, such as the following, to get started and see what’s possible. Consider tagging @Gondola before your prompts for the best results.
- Fetch my travel context and tell me who I am as a frequent traveler. What’s my favorite seat on an airplane, and how should I optimize my loyalty travel for the back half of this year? Should I be thinking about any status challenges?
- Are there any upcoming trips on my calendar that I could save money on, earn more points or rebook more effectively?
- Find me the best hotel in London for a stay from September 10 to 12. Compare member cash rates, points, and any free-night awards I can use, and recommend the best options based on my loyalty status, transferable points, and the hotels I’ve enjoyed staying at in the past.
I was frankly surprised by the usefulness and specificity of the responses, as Gondola clearly incorporated information about my travel plans and preferences. For example, here’s the start of the response to a prompt similar to the first one above.
And here’s the start of the response to a prompt similar to the second prompt above.
If you try out these prompts yourself, you’ll likely get very different responses. And there are notable shortcomings in both responses:
- The response to the first prompt said I have 58 hotel loyalty program accounts, but 58 refers to all the loyalty accounts Gondola sees for me.
- The response to the second prompt fails to account for the fact that, based on my travel history, I’m unlikely to pay that much per night for a hotel.
Gondola says its travel profile can incorporate preferences it learns over time. To this point, I used the Gondola MCP in ChatGPT for a week and explicitly told it my hotel preferences in prompts while planning actual trips. As I gave it more details of what I typically like, it began to provide more useful results.
For example, when I asked the third prompt again a week later, it gave an answer better suited to my travel style and preferences.
The detail and depth of the answers to these prompts will depend on how much you’ve previously shared with Gondola. You’ll likely get more personalized results if you’ve connected the email accounts you use for travel or otherwise given Gondola more context. But doing so also means sharing more of your travel information with the service.
Related: I tested Seats.aero’s new AI redemption tool to see if it can maximize the value of your points and miles
Bottom line
It’s easy to get started with Gondola and its MCP integration. So, especially if you’re already using a generative AI tool like ChatGPT to help plan travel or loyalty strategies, the Gondola MCP is worth trying.
There is a tradeoff, though. The more information Gondola has about your loyalty accounts, travel history and preferences, the more personalized its recommendations can be. But you have to be comfortable making more of your travel information available to Gondola and, through the integration, your AI client. And as my testing showed, having that context doesn’t mean every recommendation or conclusion will be right.
This article was the push I needed to start using Gondola’s MCP integration. Despite its occasional misses, I plan to continue using it alongside other tools as I research and plan trips.