Each year, 68 million passengers travel through here. As one of the world’s best airports, it is constantly a challenge for the airport to keep up with the expectation of their customers as they deliver services on every touch points better than the previous year.

Here are some of the objectives we set at the start of the project:

Through design research, a set of design principles and conversation guidelines was developed for the creation of the virtual assistant. Traveller profiles that we created were used to educate the clients on the customers' expectations, discovery of painpoints and opportunities. Lastly, the project ended with a feature prioritisation session with the main stakeholders that will lay out the build plan for the virtual assistant.

From January to mid-February, designers from Fjord met with 20 CAG stakeholders and spoke to some 70 travellers to understand the jobs to be done for a successful chatbot.
‍
In the first few weeks of the project, we went on the ground to gather as much information as we could. We conducted many sessions of stakeholder interviews, shadowed and observed ground information counter staff including call centre, guerilla interviews at landside and airside of the terminal and held an internal stakeholder workshop to share our initial findings.

While we found out a list of issues and queries that people approach to the information counter for, we also noticed that some issues have been consistently surfaced by travellers of all groups across channels. Among these, we sieved out issues that could be automated and self-served by travellers online and focused on such topics moving forward. Some examples are the airport's membership program, its e-commerce platform and common lost & found items. 

Through our synthesis sessions, topics that surfaced were grouped into two buckets: 1) experience at the airport and 2) chatbot-related expectations and experiences. Here are two examples within each bucket that we have gathered:

Travel-related: People are missing out on a lot of the airport
Even the most frequent travelers can’t keep up with what’s new at the humongous airport. The greater the routine, the greater their blindness to attractions and facilities that could improve their travel experience.
‍
Chatbot-related: Good customer service remembers
Repeating your case 10 times to customer support is no fun. The same is true for bots. People expect computers to have a long memory and offer a personalized service.

The next part of the project was spent turning insights into actions, which are the design principles. We also incorporated the airport's service DNA into what we think the tone of voice of the chatbot should be.

1. Sorry seems to be the hardest word to ignore
‍Fight the instinct to apologise for not being able to do something; customer support pros know it doesn’t make anything better. Nobody remembers a robot crocodile’s tears, only the outcome.
‍
2. The safe word is “Hello world” 
‍Avoid frustrating users by setting out the ground rules from the start. How should they talk? What can they ask for? At the end of every action, offer something else you can do appropriate to the moment.

Mid-way through the project, a co-creation workshop with the client was conducted. The objective was to present our findings from the discovery phase and to gather feedback.
‍
This session gathered airport staff of all seniorities and roles to share their experiences of working at the airport. While we shared our insights, every participant was invited to co-create features for the virtual assistant too.
‍
Out of several activities in this full-day workshop, there was one particular activity that stood out for me. Participants were tasked to role-play a conversation between a passenger and a chatbot. The objective of it was to put them in the shoes of a virtual assistant while representing the world's best airport. It was definitely great fun to hear how people responded to difficult customers as a chatbot!

When we thought about the experience of a virtual assistant, we also wondered how certain replies could be presented visually using existing features from Facebook messenger’s platform instead of plain text.
‍
For example, how would a flight information card look like? How could the chatbot emphasise on different flight situations visually?
It was a chance for the team to explore innovative ideas and push the boundaries of the usual chatbot. Without hesitation, I jumped on the opportunity and started working through Facebook’s design guidelines for its messenger. Despite having a limited time on this portion of the project, I had lots of fun crafting and ideating on other possible features using the available cards in the design guide.

At the end of the project, the team successfully launched a working chatbot to the members of public across the internet together with the technology team from Accenture. This is accompanied with an updated version of a customer journey map as well as a backlog of features to be further built and validated with the public.

In this top-heavy lean team, I mainly played the role of a service designer with a supporting role of an interaction designer. Besides meeting and presenting to the clients, the two design leads and I worked together from research and workshop planning to ideation and execution. Post workshop, while one of the design leads took on the role of a writer to document our research, I spent most of the time making hypothetical messenger chat conversations wireframes used for user testing later on as well as documenting our findings visually into a journey map.

After I got a diploma in Aviation Management, I worked in the aviation sector for two years. Naturally, I was very excited to be on this project as I was familiar with this industry. In the earlier phase of this project, I tried my best to explain different aviation terminologies to my teammates and shared my experiences whenever appropriate. That gave me a lot of self-confidence and I knew that with the right team members, this project will have a very good outcome.
‍
I also appreciated that we were working at the client office daily as we had direct access to the passengers whenever we needed to conduct an impromptu guerrilla interview or validate a feature on the virtual assistant. Our client trusted us and gave us plenty of freedom to explore the project, and provided us with support readily whenever needed. All of these added up as ingredients for the success of this project.