Making it faster and more trustworthy for remote workers to find and judge Airbnb listings that better support their work.
Remote workers can filter Airbnb for Wi-Fi and a workspace, but listings don't show whether either is good enough for work. So they open listings one by one, search the photos, read reviews and message hosts, on top of checking their daily-living needs. It takes a lot of time, and it's still hard to trust what they find.
Based on interviews with 4 remote workers, I designed a way to pin up to two priority filters, such as Wi-Fi and Workspace, so their details show up right on each result card and listing page. I tested it with real users and in an AI-simulated pilot, and improved it after each round.
Many remote workers now travel while they keep working, so a place to stay has to support both their work and their daily life. From my own trips, and from friends' complaints, I knew how hard it is to tell before booking whether a place will really work for working. I wanted to explore how Airbnb could support these travellers better.
Remote workers are a big group of Airbnb guests, and Airbnb is already investing in them. But when they search, there is still no easy way to judge the Wi-Fi and the workspace.
Remote workers who use Airbnb to stay somewhere while they keep working may struggle to find and evaluate listings that meet their work needs, such as reliable internet and a suitable workspace. This leads to uncertainty before booking and can cause problems with working during their stay.
The main goal of the research was to validate my hypothesis and to uncover the pain points remote workers face with the current Airbnb experience, so I could find the areas where design could help. To do that, I followed the remote worker's journey and wanted to understand four moments:
How remote workers search for listings that can support their work.
How they judge whether a listing meets their work needs, such as Wi-Fi quality and the workspace.
What makes them feel uncertain or confident before booking.
What problems they face with working during the stay.
To check that the problem goes beyond my own participants, I looked at Reddit posts from remote workers who had booked Airbnbs for work.
6 Reddit posts · 4 frustrations
“The listing said ‘fast WiFi’… In reality the connection drops several times a day… Video calls are a gamble.”
“Multiple amenities weren't there (e.g. dedicated work space)… The property photos must have been very old.”
“It just takes ages to trawl through Airbnb until I finally find somewhere with an actual workspace, then just to find out the wifi is crap.”
“I've had to run to cafes when something important comes up.”
One remote worker even built a free Chrome extension to find Airbnbs equipped for remote work. The need is real, and Airbnb doesn't meet it today.
Airbnb's own guide for hosts says guests may look at the photos before reading the description. This gave me an early idea: a separate photo album just for the workspace.
How I prepared the questions:
I'm a UX designer working on a self-initiated portfolio case study about Airbnb. I'm at the user research stage and I need help writing questions for semi-structured user interviews.
Background and motivation:
Remote work has become a normal way of working for many people, and some of them travel and stay in different places while they keep working. For them, a place to stay has to support both their daily life and their work. Airbnb already offers filters for amenities like Wi-Fi and a dedicated workspace, but from my own experience, and from complaints I kept hearing from friends, it is hard to judge the quality of these amenities and to trust that they will really meet work needs before booking.
Hypothesis:
Remote workers who use Airbnb to stay somewhere while they keep working may struggle to find and evaluate listings that meet their work needs, such as reliable internet and a suitable workspace. This leads to uncertainty before booking and can cause problems with working during their stay.
Participants:
People who have worked remotely while staying at an Airbnb at least once. The length of the stay doesn't matter. Each interview is about 30 to 45 minutes, held online.
What I want to learn:
- Why they chose to work from another place, and the context of that trip
- How they searched for a place and which factors they considered most
- How they judged whether a place was suitable for working (especially Wi-Fi and workspace)
- What made them trust or not trust a listing, and where they felt uncertain or hesitant before booking
- Whether they used tools or sources outside Airbnb, and why
- What their experience was after arriving: did the place match their expectations, and did anything affect their ability to work
- What they wish Airbnb showed more clearly
Please write the interview questions with these rules:
1. Group them into sections that follow the journey: warm-up, searching, evaluating, booking, during the stay, wrap-up.
2. Use open-ended, non-leading questions. Don't suggest the answer or assume a problem exists (for example, don't ask "Was it hard to find good Wi-Fi?").
3. Ask about real past experiences ("Tell me about a time...") rather than hypothetical opinions.
4. Add 1 or 2 optional follow-up probes under each main question.
5. End with a short task where the participant shares their screen and walks me through how they would search for and choose a place on Airbnb for a work trip.
6. Keep the total to about 15 to 20 main questions so it fits the interview time.
7. Use simple, clear English, since some participants may not be native speakers.
Screen-share task during the interviews
After each interview, I analysed it separately with ChatGPT. I gave it the transcript and asked it to pull out every fact the participant shared, one fact per bullet point, including any condition attached to it, and staying close to their own words.
I'm a UX designer analysing a user interview for a research project about remote workers who stay at Airbnbs while they keep working. Below is the transcript of one interview, including my questions and the participant's answers. Please go through the whole interview and extract all the key points the participant shared, as a list of bullet points. Rules: 1. Each bullet point must contain only one fact. 2. If a fact depends on a condition (for example "if..., then..." or "only when..."), include that condition in the same bullet point. 3. Start each bullet point with the participant's name. 4. Stay close to what the participant actually said. Don't interpret, summarise or add anything they didn't say. 5. Include every relevant fact, even small ones. It's fine if the list is long. Examples of the format I want: - Nick prefers to message the host before booking to make sure the Wi-Fi is good enough. - For Deborah, a dining table is enough as a workspace. Interview transcript: [paste transcript here]
Next, I grouped the points into themes. I didn't treat the first themes as final: I kept them open, refined them with every new interview, and reviewed them myself at the end.
When I first asked ChatGPT to cluster the points from the first interview, the groups didn't fit what I needed. It put Wi-Fi and groceries together under “basic needs”, and some groups, like “User expectations from the platform”, were too broad to be useful.
So I changed my approach and asked it to group the points by topic and by the user's story, guided by my interview questions and the goal of the project. For every next interview, ChatGPT placed the new points into these themes and added a new one when something didn't fit. After all four interviews, I reviewed the themes myself: one was renamed, and the “Other” points were spread into two new themes.
First draft, kept open
Final themes, with the “Other” points spread into the new ones
I gave all four analyses to ChatGPT to find what came up again and again, in three steps:
I'm a UX designer working on a case study about remote workers who stay at Airbnbs while they keep working. I interviewed 4 participants. For each interview, I extracted the facts as bullet points (one fact per bullet) and grouped them into themes. My hypothesis: Remote workers who use Airbnb to stay somewhere while they keep working may struggle to find and evaluate listings that meet their work needs, such as reliable internet and a suitable workspace. This leads to uncertainty before booking and can cause problems with working during their stay. Below are the clustered bullet points from all 4 interviews. Please synthesize them: 1. Look across all interviews and find the patterns: points, behaviours, needs or problems that come up for more than one participant. Ignore points that only one person mentioned, unless they are strongly related to a pattern. 2. Group the patterns under clear, meaningful headings. Write each pattern as a short, general statement about users (for example "Users rely on listing photos to identify workspace availability"), not about one person. 3. Where participants differ, say so in the pattern (for example "Some users accept basic setups for short stays, while others need dedicated setups for longer stays"). 4. Keep it close to what participants actually said. Don't add assumptions or solutions. 5. Then summarise the patterns into a shorter list that keeps only the most important and recurring ones. 6. Finally, turn the summary into 2 to 3 key insights. Each insight should have a short title and a short paragraph that explains the user need, what is missing in the current Airbnb experience, and the consequence for users (for example extra effort, time or low confidence). Interview analyses: [paste the clustered bullet points of all 4 interviews here]
How many of the 4 participants mentioned each pattern:
The full list of patterns:
The patterns came down to two insights, one for each of the two work needs.
Airbnb gives no clear details about the workspace, so users search for a desk in photos and still have to ask the host.
“I was looking for a place that I could see photos of all the rooms and literally visually count if there was a desk.”Lucas
Listings only say Wi-Fi exists, not if it is fast or stable enough for work, so users look for proof from hosts and reviews.
“Just saying Wi-Fi to me is honestly a bit useless.”Ehsan
Hypothesis supported · 4 of 4 participants
They didn't trust the filters alone, so they checked everything by hand.
They only felt confident after all that extra effort.
Based on the two key insights, I framed the problem as one user story.
As a remote worker, I want to quickly find and evaluate Airbnb listings with reliable internet and a suitable workspace, so that I can confidently book a place where I can work effectively.
How might we help remote workers quickly find and confidently evaluate listings with reliable internet and suitable workspaces?
I first mapped the current flow to see where remote workers lose time, then designed a proposed flow.
Pain point: the pink checks repeat for every listing, and most of the time goes into digging for Wi-Fi and workspace information.
Result: the work checks move up to the search results, so checking each listing becomes much shorter.
Even though work needs were a top priority, participants didn't want strict filters that hide listings. A place with weaker work features could still win them over with a great view or a pool, and they were happy to compromise a little on work. But at the same time, they wanted to be more sure about the work features, so they could stay in control of their decision.
For pinned details to show up, hosts need to add them, such as internet speed and their workspace setup. One participant worried that not all hosts would bother.
Although this could be a little extra work for some hosts, they also have a good reason to do it: about 1 in 5 guests used Airbnb to work remotely while traveling in 2021 (Airbnb Newsroom). Hosts with good internet and a proper workspace can stand out to these guests.
First, let's look at the final solution. Then we'll walk through the process that led to it.
Remote workers can pin up to two priority filters, such as Wi-Fi and Workspace. Their details then show up right in the search results and on the listing page, without hiding any listings.
Click through the flows below. The current Airbnb flow comes first, then the new feature flow, so you can compare them or jump straight to the new feature.
Current flowHomeStart on the Airbnb homepage.
Now let's go back to the start and walk through the path that led to this final solution.
To start the prototype, I didn't make wireframes. From the beginning, I built the pages to look like the real UI, ready for testing, because with AI this was quick and easy. I gave Figma Agent screenshots of every page in our flow on Airbnb, and it rebuilt the starting flow to look just like Airbnb.
My idea has three main pages: the Filters window, the search results and the listing page. I made a very quick sketch of each one in Figma, then asked Figma Agent to design the UI from them.
I wasn't happy with some parts of what Figma Agent designed. I changed some of them by hand and some with Figma Agent's help, until the first version was ready for the first usability test.
This was the first clickable flow of the feature. Before testing with users, I reviewed it with a developer to check the technical side.
Based on the developer's feedback, I made these changes:
This is the version I took to the first usability test. Along with the developer's feedback, I also changed how users choose their priorities.
I tested the design with three participants. Each one got the same task and was asked to think out loud during the whole process:
“Find a place in Berlin for a stay in October. Wi-Fi and a workspace are your main needs. Show me how easily you can find this information, and how much you can trust it.”
They also liked the listing page, where the workspace photos had their own album and the Wi-Fi details were at the top.
Everyone liked having the feature, but it wasn't smooth to use yet. Two things needed to change: the cards had to be lighter, and the feature had to be easier to discover.
Based on my research and the test findings, I redesigned the feature with the help of Claude.
Claude's first versions didn't always meet basic UI principles. In some places the spacing broke proximity, so related items didn't read as a group, and some parts didn't feel neat and clean. Other parts didn't match Airbnb's visual language. I reviewed each screen, asked Claude for specific changes, and kept refining until I reached the final result.
For the Wi-Fi details, I used Airbnb's own Wi-Fi speed categories, which are based on download speed. Airbnb describes each category with a sentence about what the connection is good for. On the listing cards, I used a short version of these sentences instead of category names like “Snappy”, because people understand “good for video calls” more easily.
As in Airbnb's categories, only listings with 50+ Mbps are called “Fast wifi”. The upload speed is shown next to the label, because it matters for video calls too.
To test the redesign, I ran an AI-simulated usability pilot. I built 5 AI personas from my 4 interviews, and each one did the same task twice: once in the current Airbnb flow, and once in my redesign. These were simulated users, not real people, so I treat the results as a first signal, not as proof.
Small, crossed out: current Airbnb flow. Large: redesign.
All 5 personas found and used the pin, and none of them tried to pin Kitchen, which wasn't meant to be pinnable.
Still open: one tap pinned both filters, which confused 3 personas. 3 missed the pink pin box at first, and for 4 of them a speed in Mbps meant very little.
To see the pilot sessions themselves, open the session replays. Each one shows every screen the persona saw, step by step, with their think-aloud notes.
If the feature launched, I would track:
Thanks for reading this far. If you'd like to talk about this project, I'd love to hear from you.