I spent about six years getting good at writing Airbnb listings. Punchy first line, the walkability sentence, a photo caption that tells you where the coffee is. Some of that still helps. Most of it now runs through a filter that doesn't read prose the way a person does.
Airbnb announced its summer release on May 20 with something north of 220 changes. The one that matters for a South Bay host isn't a new feature you can turn on. It's that the platform now decides which parts of your listing each individual guest sees.
What changed in May
AI Listing Highlights means two guests looking at the same property see different versions of it. A family gets your safety features and the bedroom configuration. A contractor on a six-week assignment gets the desk, the wifi speed, and the parking. Same listing, different front page, chosen by the platform based on what it thinks that guest wants.
AI Review Highlights does the same to your reviews. Instead of showing the most recent ones, Airbnb scans your whole review history and surfaces patterns relevant to that guest. Someone who cares about check-in gets your check-in reviews, whether they're your best ones or not.
Then there's Ask About This Home, a Q&A box where guests ask questions and the AI answers by reading your listing, your photos, and outside data like maps. It answers whether or not you've supplied the information. If you haven't, it answers from what it can infer, which is a polite way of saying it guesses.
Why your amenity checkboxes now matter more than your description
Here's the practical consequence and it took me a while to fully accept it.
If your place has gigabit fiber and you wrote "blazing fast internet, perfect for remote work" in the description but never tagged the wifi speed in the amenities, the AI doesn't know. It can't surface what isn't structured. Your beautiful sentence sits in a description field that a growing share of guests never scroll to, because the highlights above it answered their question first.
I went through this on four of my own listings in June. One of them, a two-bedroom in Willow Glen, had been renting fine for years. Untagged: dishwasher, EV charger, workspace with a monitor, and the fact that the parking is a private driveway rather than street. All four were mentioned in the description. None were in the structured fields. Tagging them took eleven minutes.
I can't give you a clean before-and-after number, because I changed pricing in the same window and I'm not going to pretend I ran a controlled test. What I can tell you is that the inquiries got more specific, and specific inquiries convert better than vague ones.
So the job now is boring data hygiene. Go through every amenity field and fill it. All of them, including the ones that feel too obvious to bother with. The platform can only personalize around attributes it has.
Reviews stopped being chronological
This one has a sharper edge to it.
The old model rewarded recovery. A rough patch in the spring got buried by a good summer, and by fall your listing page looked clean because reviewers saw the recent ones first. That's gone. If you've had four guests over two years mention that check-in was confusing, that's now a pattern the system can find and show to a guest who's asked about check-in, no matter how long ago those reviews landed.
Which means the fix has to be operational rather than cosmetic. You can't out-review a recurring complaint anymore. You have to stop causing it.
Look at your reviews for repeated words rather than repeated star ratings. Three mentions of "a bit noisy" across eighteen months is a pattern the AI will find even though your average is 4.9. Same with "took a while to hear back," which is the most common one I see on self-managed South Bay listings, and the easiest to fix. I've written before about how to handle a bad review when you get one, but this changes the calculus: the response matters less than it used to and the underlying pattern matters more.
The October problem
Why bother with any of this in August, when you're busy?
Because of the seasonal shape of this market. Per AirROI's San Jose data, occupancy runs around 76% in August and falls to roughly 42% by October. August hides a lot of sins. When half the calendar is empty and every host in Santa Clara County is competing for the same business traveler, small visibility differences turn into real revenue differences.
The work you do on your listing in August pays out in October and November, which is the same reason I push the fall pricing setup early rather than waiting for the calendar to go quiet and then reacting to it.
What I'd fix this week
Fill every amenity field, then go back and check the ones you skipped because they seemed minor. Wifi speed, parking type, workspace details, kitchen equipment, anything about accessibility.
Look at your photos with the Q&A tool in mind. The AI reads images now. A photo of the street parking with no caption is ambiguous to a system trying to answer "is there parking?" A photo where the driveway is obvious is not.
Read your last twenty reviews and write down every complaint that appears more than once. That list is your operational to-do list for the fall, and it's more useful than your star average.
Then fix your response time if it's slipping. Consistency is visible to the ranking system now in a way it wasn't, and response time is the piece most self-managing hosts let slide in a busy month.
Where this leaves copywriting
Not dead, just demoted.
Your title and your first photo still do the work of getting someone to stop scrolling, and no amount of structured data saves a listing with bad photos. What's changed is that everything after that first impression is increasingly mediated. The description is now more of a backstop than a sales pitch.
The other thing worth saying: this is one more argument for owning some of your demand directly. The more of the guest relationship the platform mediates, the more valuable it is to have people who book with you directly and never meet the algorithm at all. I'm not suggesting you leave Airbnb. I'm suggesting that a channel where an AI decides who sees you is a channel worth diversifying away from at the margin.
Airbnb has said an AI comparison feature is coming later this year, letting guests put listings side by side with machine-written summaries. I don't know yet how much that will move bookings. But whatever it ends up looking like, the summary gets built out of your structured data, which puts it back on the same eleven minutes of unglamorous work.
Want someone to run the listing audit for you? Get a free rental analysis and I'll go through your amenity fields, review patterns, and fall pricing before the slow season hits. Or call me at (408) 813-8001.
Sources
- Airbnb's 2026 Summer Release Changes How Listings Get Shown to Guests — The Host Report
- Airbnb's 2026 Summer Release: What Hosts Need to Know and Do — Hostaway
- Airbnb AI Strategy in 2026: What the Summer Release Confirms — Rental Scale-Up
- San Jose, California Airbnb Data 2026 — AirROI
- San Jose, CA Airbnb & Short-Term Rental Data — AirDNA
- Airbnb 2026 Summer Release — Airbnb Newsroom
Ready to optimize your property?
Get a professional evaluation of your property's STR revenue potential, completely free.
Start Partnership