Welcome to Lorraine Liu's space

Problem Space - Create Best-In-Class Virtual Assistant

As part of OMNI-Channel service strategy and a supplement to corporate websites / customer self-serve portal / Rep oriented sales&sales platform, Virtual Assistant appears to us as a new problem space.

This is a brand-new build that started in 2019 and will be a cross-year effort, including early experimental MVP scope design, research, interaction pattern building, content strategy, features development and product branding design etc.

Our goal is to build a best-in-call Chatbot format Virtual Assistant to lower live agent deflection rate and improve overall customer satisfaction.

Project Outcome

I led the beginning-to-end design journey, from the initial scoping discussion, interaction pattern defining, product position steering, intent decision tree, content model, wireframing to final mocks.

Through my design leadership, Virtual Assistant has grown into a product with established experience framework, cohesive design pattern, well-praised visual style and most importantly a solid UX foundation to develop features sustainably.

Research

Online Study

To enable ourselves to start with a wide view of what the market is offering. We in partner with Product Team to study AI Chatbot applications in Telco industry, and gets to know the trend of AI chatbot and its variants of serviceability levels:

VA tpes — Fig.1 - Virtual Assistant types based on serviceability

And for experience design exploration, my team and I looked into dozens of conversation UI design and online assistant design to inspire ourselves on how to digitalize natural language conversation.

Experience Metrics

Apart from the general User Experience criteria that we will practice towards, I also advised designers to align with the product metrics that come from Product Team to have a universal target:

- Chat Deflection Rate: percentage of the total number of chats that were resolved by Virtual Assistant;
- Helpfulness Rate: customer rating of helpfulness on individual resolutions from Virtual Assistant.

Intents-based Micro Jouney Mapping

The key design activity when designing for conversational UX is to streamline the intent-based user journeys (in this context, instead of cross-timeline journey that we usually would focus on, we actually refer to micro journey for each type of intents we support, starting from the moment customers are greeted by Virtual Assistant till the particular intent is satisfied), where my team needs to factor in:

- Service availability
- Ideal customer experience
- Conversation depth/efficiency
- Journey framework from the vendor

Micro Journey example — Fig.3 - Micro Journey Example

Interaction Pattern

The interaction between the customer and Chatbot requires a set of interaction patterns to enable: customer inquiry, informational contents, interactive contents, multimedia contents, transactional instructions, system notifications.

We started with a very generic pattern framework to validate if we can find the universal patterns across the first 20+ intent journeys. After, I advised the designers to create a high-fidelity set in detailed contexts.