Inquiries to the front desk down about 90%: how AI took over routing guests between two brands

Inquiries reaching the front desk at "WASUKI BASE Biwako" piled up on the phone and the official site's contact form, and only the manager and staff members could handle them.

After adopting NEWT Chat, inquiries to the front desk fell by around 90%. The time that freed up now goes toward looking after each guest individually.

We spoke with Natsumi Inamine, who handles sales and marketing for "WASUKI BASE Biwako," about why they adopted it and what changed on the ground. (Interviewer: Watanabe, Reiwa Travel)

Running two distinctly different brands, one for small groups and one for larger ones

— Could you start by telling us about your company's structure, and your own role? We'd also love to hear how the two brands, VILLA STYLE and GUEST HOUSE, differ.

Inamine: We run two brands: GUEST HOUSE, built mainly around group stays, and VILLA STYLE, our glamping and cottage brand. GUEST HOUSE is aimed at groups — university clubs on a trip, retreats — and can take bookings of 13 or more guests. VILLA STYLE has cottages that sleep up to 15 and glamping domes that sleep up to 10, so the range of guests is wide, from couples and families to small groups of friends.

Inquiries piled up on the phone and the contact form, and only the manager and staff could handle them

— What challenges did you face with inquiries before adopting NEWT Chat? We'd love to hear about your team, too.

Inamine: Before adopting it, inquiries came mainly through the phone and the official site's contact form, which forwards to email. This was handled 100% on site, and only the manager or staff members could handle it — part-time staff were kept focused on work like cleaning.

Inquiries tended to pile up right when staff were busy with cleaning and check-in, which made it hard to find time to sit down and reply. In the end, staff ended up answering early in the morning or in whatever gaps they could find, and there were times we couldn't reply quickly, which was a real problem.

Adopted casually, just to try it for free. Design concerns were resolved from the admin screen

— What led you to adopt NEWT Chat, and what concerns did you have going in?

Inamine: What triggered it was really just wanting to try it out for free. Since on-site staff were handling 100% of the workload, we figured it was worth trying if it could ease even a little of that, so we decided to try it on a trial basis.

Our concerns were things like whether it would break the design once embedded into our existing website — that was about it. Once we actually got into the admin screen, we found we could adjust things like the button's position bit by bit, and that resolved it.

Answers more solid than we expected. An admin screen simple enough to figure out right away

— What did you think of the accuracy of the answers and how easy the admin screen was to use, once you actually tried it?

Inamine: On the answers, we were surprised by how solid the writing was, well beyond what we expected. The admin screen was quite simple too — simple enough that even the first time, we could figure out how to use it right away.

Guests routed to GUEST HOUSE or VILLA STYLE by group size — the routing the AI learned on its own

— We heard there's an interesting way you're using this that only makes sense because you run two brands. Could you tell us more?

Inamine: Since we have two properties, we were able to train the AI so that, say, if a large group or a party inquiry comes in for VILLA STYLE, it can tell the guest that GUEST HOUSE also takes group bookings at that size. We actually put a line in the Q&A saying GUEST HOUSE is a good fit for 13 or more guests, and sure enough, it started recommending it that way.

The other direction works too: if a small inquiry — three or four guests — comes in for GUEST HOUSE, we've trained it to suggest VILLA STYLE next door as well. Having the AI route guests between our own properties was honestly a use case we hadn't quite expected ourselves.

Casual questions become a source of marketing insight

— Is there anything you noticed that's unique to a chatbot, or anything you've done to make the most of daily operations?

Inamine: On top of the information on our official site and details about the property, we've trained the chatbot on check-in and check-out times, our cancellation policy for typhoons, questions about barbecues and pets, and even nearby tourist spots. When it's asked where a guest is coming from, it goes as far as suggesting a route tailored to that area, which surprised us.

With a contact form, guests seem to sense it's "going to a person," so their wording tends to be a bit more formal. With the chatbot, though, guests send us casual questions — is there a rice cooker, what's in the barbecue set — far more freely, so we're now seeing our guests' real voice and small needs more than ever before.

We take the questions that come up often and move them higher in the FAQ, and reflect them in listings on other booking sites too. Inquiries tend to spike on weekends and holidays, so early in the week I go through the conversation log and add anything the chatbot couldn't fully answer to the Q&A. When a topic gets asked about too much, we treat it as a sign to reconsider the plan itself — voices we'd never have caught without the chatbot are feeding directly into our marketing decisions.

Inquiries to the front desk down by around 90%, with the biggest impact felt during the busy summer season

— What changed after adoption? We'd love to hear the numbers, too.

Inamine: Across the board, inquiries to the front desk fell by around 90% after we introduced the chatbot. I was surprised myself. Once we actually looked into it, it turned out there had been a fair number of phone inquiries, right down to detailed questions like check-in and check-out times, and that's what settled down significantly. Having inquiries drop by around 90% right during the busy summer season was a welcome change for me personally.

We've deliberately set things up so the chatbot doesn't answer questions that need an on-the-spot judgment call — like whether guests can bring in a watermelon for suikawari or an inflatable pool — and routes those to staff instead. That means what actually reaches the front desk now is only the content that genuinely needs an individual judgment.

Decided to continue with an annual plan — what started as a trial run has paid off

— Finally, could you tell us your plans going forward, and your honest thoughts looking back on this adoption?

Inamine: We're genuinely glad that adopting NEWT Chat has significantly eased the burden of handling inquiries. We wanted to keep it going if we could, so we've decided to continue on an annual plan.

We started by rolling this out to WASUKI BASE as a pilot property, and now that we're seeing the results, we're looking at rolling it out to the other properties we're involved with too.

— AI bridging two brands, and inquiries to the front desk down significantly too. It's been a genuine pleasure to hear this. Thank you so much for your time!