FINTECH · 2017
Designing conversation experiences for Australia's largest bank. Commonwealth Bank of Australia (Commbank) has formed a digital conversations team to push the boundaries of what a chatbot can do on mobile devices.
Techniques
Heuristic Evaluation, Persona Creation, Sitemap, User Journeys, Wireframe, Prototyping & Usability Testing.
My Role
Led the design sprint for transaction enquiry conversation flows, build high fidelity prototypes to test with customers and communicate solutions to stakeholders and scrum teams.
Goal
To design a bespoke conversational experience on a new virtual AI chat platform called 'Ceba'.
Tools Used
Sketch, Axure, Keynotes
Devices
iOS & Android
Ceba (see-bah) is a virtual banking assistant who can answer a range of day-to-day banking questions and provide users with helpful info. A sophisticated chatbot trained to give users in-the-moment, digital support. Not to be confused with the bank's messaging or live chat support, where users have an online conversation with one of CommBank's humans.
About 45% of our customers tap on their transactions every day, of these, 37% are calling about issues relating to their transactions.
By providing relevant information, we can help customers relate and recall how, when, and where the transaction was made. Flushing out these pathways allowed the team to understand potential use cases and conversation scenarios.
Validation
Success
The discoverability of Ceba was easy and all users clearly articulated where the purchase was made using suggested names. They also articulated that the information helped them recall what the transaction was about.
Considerations
Some would prefer to receive all the information about their transaction at once to avoid tapping on 'Tell me more' multiple times.
After my involvement in the project, the team delivered the initial MVP flow in 2019. The initial results were positive:
22% increase in virtual chat engagement and usage.
17% reduction in calls related to transaction inquiries.
Think big and start small to shorten experimentation cycles.
Provide guided responses to help users shorten their conversation flow.
Get feedback early and iterate fast.