Booking Assistant case study
Answers in minutes, at any hour
A chatbot that answered guests’ questions about their stay, and handed the rest to a person.
I set the vision for Booking Assistant and led design across Booking.com’s messaging platform as it grew from one team to seven.
Questions went out, few answers came back
properties actively replied to guest messages
Booking.com’s messaging was basic and few properties replied. Guests went without answers, partners saw it in their reviews, and customer service costs kept rising.
Designed for four groups at once
- Guests
- Anyone who books a place to stay on Booking.com
- Partners
- From large hotel chains to single apartments
- Customer service agents
- Answering calls and messages at all hours
- Third parties
- Partner APIs, and channels like Facebook Messenger
The vision: fast answers for guests, time saved for partners, and lower customer service costs for Booking.com.
A bot that knows when to hand over
We piloted with ten customer service agents in Amsterdam, who sorted every incoming message by topic. That taught the Assistant to answer routine questions instantly. Anything else went to a person: an agent, or the property.
Time back for every property partner
Dear James,
We are excited to welcome you to Relax Hotel for a total of 1 night. If there’s anything you need, please don’t hesitate to reach out to us in advance of your stay.
Best regards,
Relax Hotel
I’d like to request an early check‑in at 11:00–12:00. Is that ok?
Automatically approved- One inbox for every guest conversation
- Automatic replies that partners could customise
- Templates written once, then sent for them
- Quick replies and monitoring dashboards
Built to grow past a single team
designers on the team, from 2016 to 2018
Every messaging screen was systemised and built as components in code with engineering leads, so we could move fast across channels. I hired with our recruitment team as the platform grew from one team to seven.
What it added up to
Of stay-related questions were answered automatically
To an automatic answer
Saved a year in customer service costs
What I learned along the way
- Anticipate resourcing needs
- Talk to users, regularly
- Design systems provide velocity in the long run
Next project
Shop app
The shopping app that reached #1 in the App Store’s shopping chart
View case study →