Hostaway asked 700 operators how they use AI. Only one in five has it doing actual work

Nearly everyone in short-term rental management is now using AI for something. Almost nobody has it running unsupervised. That is the short version of Hostaway's 2026 STR AI Benchmark Report, published 2 October from a survey of more than 700 operators across 62 countries.

The five groups

Hostaway split respondents by how far they have actually taken it.

Twelve percent do not use AI at all. Twenty-six percent use it occasionally, for quick answers. Thirty-two percent, the largest group, use it as a thinking partner. Twenty percent run AI agents or automated workflows. And ten percent say AI is in use somewhere in their business but are not sure how it fits together.

Add the first four and you get 88% reporting at least some use, up from 61% in Hostaway's earlier 2026 report. That is a fast climb. But the number that matters operationally is the 20% running agents and workflows, because that is the only group where AI is doing work rather than helping someone think.

Scale drives it, experience does not

The clearest split in the data is portfolio size. Operators with 50 or more properties are eight times as likely to use automation as single-property managers.

Tenure barely registers. Automation sits between 19% and 24% for anyone who has been operating more than a year, regardless of how long that year count runs. Being in the business for a decade does not make you more automated than being in it for two years. Having fifty properties does.

That tracks with how this category tends to work. Automation gets bought when manual process breaks, and manual process breaks at a property count, not at an anniversary.

The payoff claim, and how to read it

Hostaway reports that automated operators are about five times more likely to save 11 or more hours a week and $1,000 or more a month than peers who have not automated.

Worth being careful with that one. This is a self-reported saving, from a survey run by a PMS vendor that sells automation, and the causation runs both ways: the operators with fifty properties are both the most likely to automate and the most likely to have large absolute savings available to find. The number is a reasonable signal that automation pays at scale. It is not evidence that a three-property host will free up eleven hours.

What is actually holding people back

The trust numbers are the most useful part of the report for anyone selling or buying this.

Among operators who have not automated, 37% want to review everything before AI acts. Among those who have, 74% are comfortable letting AI handle small tasks unsupervised. So the barrier is not a view that AI cannot do the job. It is the step of letting it act without a human in the loop, and that resistance drops sharply once people have watched it work.

Both groups name the same weaknesses: complex or unusual situations, reading guest tone and nuance, and general accuracy.

Those three are not solved by better onboarding. They are the real edges of the technology, and they explain why nearly every PMS shipping AI features this year, Hostaway's own CoHost included, has settled on a prepare-and-approve pattern rather than letting the assistant send.

What it means for operators

If you are below about twenty properties and feel behind, this report says you are not. The thinking-partner group is the biggest one and the automated group is one in five.

If you are above fifty and still doing guest messaging by hand, you are now the outlier in your own size bracket, and the competitors who automated are claiming double-digit hours back each week.

The report was written by Rikaza Marie and is published on the Hostaway blog.