RSU by PriceLabs ran a live debate on June 24, 2026, on a question more property managers face every month: build your own AI in vacation rentals business, or wait for your software to add it. Boostly founder Mark Simpson argued for building. RSU founder Thibault Masson, who also heads product marketing at PriceLabs, argued for waiting. Before either made their case, the audience was asked to vote.
Most Property Managers Said They Want to Build
When Simpson asked people to type “build” or “wait” in the chat, the replies ran almost entirely to build.
The choice shows up in two ways:
- The AI comes to your software. Your PMS or pricing tool adds AI inside the product, and you use it there. Nothing to set up.
- The software comes to your AI. You make ChatGPT or Claude your home base and connect your tools to it through an MCP, a standard plug that needs no code.
Most leaned toward the second path. The rest of the hour examined whether that instinct holds up.
Cost Isn’t What’s Stopping Operators
The clearest finding of the session came from a live poll. Asked what stops them from building their own AI workflows, attendees said:
The wall is knowledge and time, not budget. Asked where they would point AI first, 39% chose pricing and revenue management and 36% chose guest messaging. RSU has covered how the strongest operators use AI to read their pricing data while keeping a human on the final number, in this report from Scale UK 2026.
The Case for Waiting: Your Vendors Are Already Building This
This was the half the room kept talking over, and it deserves a fair hearing.
Masson’s points, in his own framing:
- The best tool is the one you actually use. Your PMS, your revenue tool, and others are all rolling out AI inside their products in 2026. If you are busy running a business, using that is a reasonable choice, not a lazy one.
- Why rebuild what a full-time team already owns? The thrust of his case: a weekend project cannot match software that a dedicated team has maintained for years, especially the parts that have to keep working every day.
- Embedded AI comes with guardrails. A general tool like Claude is powerful, but it does not carry the operational rules and safety limits your software has built up. A vendor’s AI is less likely to do something costly because the product already knows how your operation runs.
- The costs are still ahead of most operators. Only 5% named cost as a barrier, but most have not started building yet, so they have not met the bill: connector and usage fees, plus real money spent on mistakes. Masson’s advice was blunt: ask your vendor whether they have a connector and what it costs before you assume building is cheaper.
Simpson’s side is real too. Operators who move first get a head start, some are cutting their software bills by building tools themselves, and an employee who leads on this can make themselves hard to replace. The bigger companies tend to land on the cautious side: Sykes Cottages and Casago have explained why they more often buy than build, and why their best wins came from internal tools rather than flashy guest-facing ones.
The Line That Settles It: Build on Top, Not Underneath
When an attendee asked whether they could just build their own PMS and drop the fee, both speakers said no, for the same reason.
A PMS holds the channel-manager link to Airbnb, Booking.com, and Vrbo. That link runs two ways and can never drop, or you get double bookings and missed reservations. A builder tool can produce something that looks finished while missing the last 5% to 10% that actually plugs into your systems.
So the rule is simple. Build AI tools on top of your PMS. Do not try to replace it. RSU made the same point in its breakdown of where Airbnb’s own AI helps and where it won’t: native AI raises the floor for beginners, but it does not replace the specialist tools a professional relies on. The question to take to your vendor is the one Masson kept returning to: are you building a connector, and what will it cost?
RSU’s Read: Decide on Purpose
Both paths are valid. A closing poll showed about half the audience planned to do a mix of both, and that is a fine place to land. The mistake is drifting into it by accident instead of choosing it.
If you build, build on top. If you wait, push your vendor for a timeline and a connector price, and expect real answers through the autumn, when most will have shipped something. The good property management systems will go all in on connectors and stand out. The ones that do not will fade.
The operators who lose this year will not be the ones who waited or the ones who built. They will be the ones who never decided which they were doing.
