AI at VA
Discovery / Research / UX
Virgin Australia
Overview
Like in all industries and parts of the world, AI is a hot topic right now, and it’s no different here at Virgin Australia.
Over the last 18 months, I’ve been fortunate enough to work on a few varying AI-based projects, all with the focus of satisfying slightly differing needs and use cases, but all ultimately aiming to provide new and exciting features for our customers.
Shopping: ‘AI Trip Planner’
Through multiple pieces of user feedback and conducted user research, it was clear that while shopping capabilities are becoming more prevalent across all industries, users still had that ‘trust’ issue when it came to converting on products like flights.
Although guests appear to have a general lack of trust when it comes to AI currently, with reservations about the accuracy of the information being provided, it was however clear that guests would be open to having some functionality in the future (research conducted over 12 months ago - 2024/25).
Early stage user journeys
Shopping: ‘ChatGPT VA App’
The problem
Virgin Australia business highlighted the need to leverage Generative Artificial Intelligence (GenAI) within its digital infrastructure to not only improve internal efficiencies, but create revenue opportunities and uplift the guest experience via support channels.
By leveraging AI we hope to allow guests to achieve their travel and planning goals through an interactive experience which will increase conversion.
Guests also highlight through feedback and NPS there’s a lack of efficient help and support via online channels, specifically via chatbot functionality.
This project summary will be split into two key areas of AI focus - Shopping and Channel Support.
Discovery research
In partnership with Sabre, we set out on designing and building a travel exploration chatbot from scratch, one that specifically serviced the needs of our users who wanted to have that playful, interactive style bot that would serve up new destinations, while also offering the guest real-time flight options with up-to-date pricing.
After initial flows, requirements, use cases and identified needs were established, we created wireframes and mocked up ‘chat flows’ that we could use to kick off the build and test internally.
AI Trip Planner - Production
Original mock ups, used for internal testing
Over a period of many months and after multiple rounds of testing, we were able to test our production version of the AI Trip Planner on dedicated pages which we thought best suited to house the bot which were based more around exploration and destination discovery.
Today, after being live for over 6 months, the AI Trip Planner has recorded over 30,000 unique visits.
After many learnings and discovery off the back of the Trip Planner, it was clear there was strong appetite for a shopping based chatbot that provided quick and responsive results for our users. As such, a recent collaboration with OpenAI has brought about the inclusion of a new Virgin Australia shopping app within the ChatGPT ecosystem.
Multiple workshops were held to understand the core use cases we wanted to target for first release of the product offering, with strict requirements in order to keep to MVP and initial release, and initial design mock ups were created prior to final UI handover and build.
The VA ChatGPT App is due for upcoming release into production in the near future.
AI workshopping
Support channels: DVA (Digital Virtual Assistant)
Initial mock ups (non-prod design).
In a similar fashion to the AI Trip Planner, we set out at trying to better service our guests during times of need, disruption and confusion by establishing a support based chatbot.
By integrating a digital virtual assistant (DVA), we can alleviate the pressure on the Guest Customer Care and Member Contact Centre teams while providing our valued guests with live assistance. This technological solution would enable our guests to resolve their queries in real-time and obtain personalised assistance, improving their overall experience with our website. Additionally, this innovation will enable the GCC/MCC team to prioritise complex queries, leading to improved efficiency and increased productivity.
Early stage UX mock ups
Since its inception last year, the DVA has serviced thousands of customers and dramatically helped reduce over-the-phone call volumes for our contact centres, enabling both our operations to work more smoothly, and allow our guests to self-serve at quicker pace.
Outcome
AI continues to be a huge learning curve across all industries, not just in aviation, and while the implementation of these products has mostly been positive, there are clearly learnings that can be taken into future iterations and decisions to better serve not just the business, but ultimately the consumer.
From where we are today, compared to a year or more ago, we have made some clear advancements in the AI space in order to better serve our customers, but with the continual growth in the area, AI will need continued focus, investment and user consideration to ensure it is implemented in the most applicable way it can be.
Key roles & responsibilities
_Ideation Workshops
_Product Design
_User flows and wireframing
_Prototype creation
_Collaboration with PO/Dev on build design iteration
_Stakeholder management