A lot of the conversation around AI in short-term rentals focuses on automation. Let the software handle the decisions, free up time, and focus on growth. It sounds promising, but it doesn't quite match the conversations I've been having with property managers.
What I hear more often is something closer to: AI spotted something I wouldn't have caught, and I decided what to do about it. For most property managers, the value comes from having better information at the right moment rather than handing decisions over entirely.
Until recently, an AI assistant could only work with whatever you typed into it. It had no way to see what was actually happening in your business, so its suggestions stayed generic.
That’s now changing. A standard called MCP (Model Context Protocol) allows AI assistants like ChatGPT or Claude to connect directly to the systems hosts and property managers already use. Instead of relying solely on what a user types into a prompt, they can work with reservation data, guest history, maintenance records, financial information and other operational data in real time.
As a result, AI can answer more useful questions, spot patterns that might otherwise be missed, and surface issues before they become problems. But importantly, the operator still decides what action to take.
What AI can see, and what it can't
A returning guest with two previous five-star reviews might rebook the same property for six nights. AI with access to the PMS flags it immediately: that property has three unresolved maintenance tickets. The AI raises the issue and suggests reviewing those tickets before arrival.
That recommendation is useful because it highlights something that could affect the guest experience. Deciding what happens next requires information that may never appear in the PMS.
The AI doesn't know that the owner is reluctant to approve maintenance spend. It doesn't know that two of the tickets are cosmetic while one genuinely needs immediate attention. It doesn't know that this guest has previously raised accessibility concerns that were never formally documented.
The operator knows all of that. They know which issue needs resolving first, how to approach the owner, and whether it's worth contacting the guest before arrival. The software identifies the issue, but the response still depends on the person managing it.
What operators built with access to their own data
Hospitable recently hosted the industry's first MCP ‘Buildathon’, inviting property managers and AI enthusiasts to experiment with AI agents and see what they would build once AI could work directly with PMS data.
The response highlighted just how many opportunities emerge when operators can connect AI directly to the information they work with every day.
One participant created Found Money Engine, which analyzes guest conversations, reviews, checkout forms, cleaner notes, and reservation data to identify guests most likely to return and generate personalized follow-up outreach.
Another built a Weekly Property Intelligence Report that reviews bookings, guest sentiment, listing changes, performance metrics, and guest communications across multiple properties. The tool automatically generates management reports, maintenance tickets, operational updates, and draft responses.
A third developed a Damage Claim Recovery Dashboard that aggregates portfolio losses, prioritizes claims by urgency and value, and generates structured recovery workflows designed to help operators recover lost revenue.
Other projects focused on the guest experience, including AI-powered welcome flows and guest guides that personalize communication throughout the stay while helping operators gather useful feedback and insights.
What stood out wasn't any single project. It was the range of ideas that emerged once operators had the tools, support, and confidence to build. Participants created solutions spanning revenue management, guest experience, operations, maintenance, business intelligence, and portfolio oversight. Many addressed challenges that would never appear on a traditional software roadmap because they reflected the specific needs of individual businesses.
The projects themselves were impressive, but they also highlighted an important point. AI can identify patterns, surface information, and connect data from across a business far faster than any person could. That doesn't mean it has the full picture. A tool might identify a guest who is likely to rebook. It won't necessarily know that the guest had a difficult conversation with a property owner during their last stay. A dashboard might flag a maintenance issue as low priority based on historical data. It won't know that a VIP guest is arriving tomorrow.
The real skill is knowing when to trust AI
Access to live operational data improves the quality of AI recommendations, but it doesn't remove uncertainty. As AI becomes more connected to operational systems, operators will increasingly find themselves weighing recommendations against what they already know about a guest, owner, or property.
Some recommendations will be obvious. Others will require additional context. Experience still plays an important role in deciding what action to take.
An AI assistant can analyze reservation histories, maintenance records, guest communications, and financial performance. It still won't understand every owner relationship, pick up every nuance in a conversation, or recognize when a situation calls for a phone call instead of an email.
A guest may sound frustrated long before they leave a negative review. An owner may respond poorly to a written update but react differently during a conversation. A maintenance issue that looks minor in a ticketing system may suddenly become urgent because of who is arriving next.
Those decisions still rely on experience, judgment, and familiarity with the people involved.
The operators getting the most from these tools tend to view them as a source of information rather than a source of answers. AI helps surface opportunities, risks, and patterns that might otherwise go unnoticed. The operator decides what happens next.
If you're on Hospitable's Host, Professional, or Mogul plan, you can connect an AI assistant to your account via MCP today. Find out how at hospitable.com.
.webp)
