Case Studies

The numbers, not the adjectives

Every figure on this page is counted from production, real conversations, real properties, measured the same way for each.

properties with live guest traffic: 190+properties with live guest traffic
guest conversations handled: 265K+guest conversations handled
countries: 6countries
of conversations on WhatsApp: 97%of conversations on WhatsApp

What full automation looks like

Resort

Nine hundred questions a day, and almost none of them reach a person

Hamat Gader

Northern Israel

The reservations team was answering the same handful of questions all day. Are you open. Is there availability. Can we bring the children. What does the spa package include. Every one had a known answer, and every one still cost an agent a phone call.

Going live took a month. Week one was plumbing and voice: connect SimplBooking and Optima, load the property's own knowledge, train the team. Weeks two and three were spent correcting the assistant against real conversations. By week four it was carrying the channel.

Since then guests have sent 158,407 messages and the assistant has answered 193,544 times. People stepped in 2,636 times, about one reply in seventy five. The team did not stop talking to guests. They stopped answering the same question twice.

“It simply made our work much calmer. A smart bot that answers everything about Hamat Gader, the spa and the hotel, available 24/7. Since going live we have seen a significant drop in workload and an increase in bookings.”
Hanna Ostash, Reservations Manager, Hamat Gader

Connected to SimplBooking · Optima (SilverByte)

Measured results

of replies sent by AI: 98.7%of replies sent by AI
AI messages sent: 193,544AI messages sent
guests served: 37,139guests served
replies needed a human: 1 in 75replies needed a human

Five months old and compounding

Urban boutique hotel

Enquiries grew 37%. The work of answering them grew 3%.

Margosa

Jaffa, Tel Aviv

Margosa runs click to WhatsApp campaigns, so a guest who taps an advert lands in a conversation rather than on a form. It works, which is the problem: every successful campaign arrives as more messages for a small front desk to answer.

The assistant went live in March and started quietly, on about one reply in a hundred. By August it was handling three in ten, and it climbs every month as the knowledge base absorbs the answers the team keeps typing by hand.

The number that matters is the gap between two growth rates. Between April and July guest messages rose 37%. Replies written by a person rose 3.5%. The hotel got busier and the front desk did not.

It has picked up the hotel's instincts, too. Asked whether there is a pool it says no, then in the same breath offers the sea twelve minutes away, beach towels at reception and the flea market around the corner. A twenty room enquiry goes to the team with the dates already attached.

Measured results

AI share since March: 0 to 30%AI share since March
guest messages, April to July: +37%guest messages, April to July
replies written by the team: +3.5%replies written by the team
proactive messages sent: 8,221proactive messages sent

Half the replies, none of the waiting

Glamping

17,380 answers that would otherwise have arrived in the morning

Noor Glamping

Israel

Glamping guests ask questions that only matter right now. Is the road passable. What time does the gate close. Is there power in the tent. Answered at midnight, that is worth something. The same answer at nine the next morning is worth very little.

The assistant has sent 17,380 replies, a little over half of everything guests received. The team sent the rest and is still firmly in the conversation, which is rather the point. The split is close to even and a guest cannot tell where one ends and the other begins.

What changed was not headcount. It is that nothing now sits unanswered overnight.

“By using vGuest, we're allowing our customers and our guests to get faster replies.”
Gilad, Noor Glamping

Measured results

replies sent by AI: 17,380replies sent by AI
of all replies: 53%of all replies
guests served: 4,547guests served
guest messages handled: 32,398guest messages handled

The team never left the conversation

Serviced suites

From 44% to 58% in three months, without anyone changing how they work

Sea Executive Suites

Tel Aviv

Not every property wants to hand the channel over, and Sea Executive did not. Their team answers guests directly, alongside the assistant. That is a supported way to run it, not a failure to launch.

In May the assistant sent 43.5% of replies. By August it was 58.2%. Nobody rewrote the strategy in between. The knowledge base simply absorbed the answers the team kept giving by hand, and each month it needed them a little less.

It is also the most talkative account in the portfolio: 6.9 guest messages per guest, against 4.3 at Hamat Gader. Guests hold conversations with it rather than firing questions at it. Behind that sit 2,783 proactive messages, reaching out before the guest has to ask.

Measured results

AI share, May to August: 44 to 58%AI share, May to August
messages per guest, a portfolio high: 6.9messages per guest, a portfolio high
proactive messages sent: 2,783proactive messages sent
guest messages handled: 10,189guest messages handled

How we measure

“AI share” is the proportion of replies a guest received that were sent by the assistant rather than a person, counted across every stored conversation for that property, with no sampling and no exclusions. Where a property’s team also answers guests directly from the WhatsApp Business app, those replies count as human. That is why the range across accounts is wide and honest: where vGuest carries the channel, as at Hamat Gader, the AI share reaches 98.7%; where it shares the channel with staff, as at Sea Executive Suites, it starts lower and climbs, from 44% to 58% in three months.

Book a Demo

15 minutes · No commitment

Prefer WhatsApp? Message us →