designing ALIA

Over the years, the customer support team has grown 3x more than it was at a time and is growing more to accommodate the various needs of customers as well as ensure customer satisfaction stays top, be it from as little as not receiving email communications to as large as KYC verification issues or transaction issues.

Starting with addressing the elephant in the room, we went through a backlog of customer queries, and we had sessions with various stakeholders across the product and customer support teams in their various departments to gain more insights on the problem, their limitations, and pain points. From there, we were able to identify user flows and map out the customer journey.

While it is a no-brainer as to existing chatbots in various industries, it was a bit of a challenge to find one very much related to the crypto space for analysis. However, I evaluated popular key features based on the insights we got from meeting with stakeholders and identified missing functions that ALIA could possibly utilize.

I created user personas to better empathise with the needs and goals of our target users. This helped to keep the needs of our customers at the forefront and prioritize key features.

Good design for AI goes beyond attractive interfaces. It’s about creating a seamless, intuitive experience that leverages AI capabilities while maintaining a human touch.

For ALIA, this meant designing natural conversational flows, providing clear options for customers to navigate, and ensuring transparency about the assistant's AI nature.

Before embarking on high fidelity mockups, we brainstormed potential UI designs, illustrating wireframes to get a feel for the product feature and its core functionalities.

With this flow in mind, we went on to create the digitalized mid-fidelity mockup on Figma. We went on to share this with stakeholders to get initial feedback through this iterative process.

The user flow was iterated to allow for a new addition to the existing product and to also be as intuitive as possible for customers. Following various meetings, we were able to go from iteration 1 to proposed solution designs that meet user needs and business goals.

Balancing functionality with feasibility was a crucial aspect of Alia’s design. We worked closely with the engineering team to understand the capabilities and limitations of the AI system. This collaboration helped us design within the constraints of what was technically possible while pushing the boundaries of customer experience.

One key learning was the importance of designing for edge cases in AI. For instance, we implemented features to ensure Alia could understand and respond to imperfect queries like wrong spellings and unsupported text.

Another edge case was the character count and an alert to know when you have crossed it. To get customers to be as concise as possible, we added this 160-character count. This was also important because shorter prompts meant fewer resources committed to answering them.

Designing for AI reinforced that our responsibilities as designers extend beyond just creating visually appealing interfaces. We must think about how our customers interact with the systems and the feedback they get, more like a cause and effect.

Every AI needs a helping hand; The initial launch of ALIA is just the beginning. We’ll set up analytics to track key metrics like response time reduction, customer satisfaction, and the accuracy of Alia’s responses. This data will drive our ongoing iterations and improvements.