Context:
Create a AI conversational tool to support customers in solving support issues independently, and being redirected for more complex scenarios and cases to customer service representatives.
Timeline: 2024 - present
Role: Lead design, vision and solution desing
Title: Chief Product designer
Main responsibilities: vision design, solution design, conversation design, cross team alignment, stakeholder management, mentorship
Key achievements:
- long term vision designed and scoped, in progress
- cross-team collaboration process outlined and executed, currently using for design team collaboration
- pioneered research driven decisions with 5 active research rounds, including eye-tracking sessions
- mentoring and guiding junior designers in the team
Problem statement – overall scope of the platform
As briefed for the MVP – design a customer facing banking AI assistant that would serve customers by supporting them with executing tasks in the Mobile bank solution. The AI assistant solution should be able to answer the most commonly asked questions to customer support and deflect the customer to a customer service representative when the task is too complex or sensitive to answer
My responsibilities
As a team based specialist, my responsibilities covered the full design process.
Leaning in my Chief role, I assumed ownership of shaping and leading the product vision from a customer prespective.
This meant I would start with an ideal solution, without technical restraints and break down into feasible increments for progressive implementation.
Key responsibilities and tasks
- Discovery and research
- Vision and concept design
- Solution design
- Key functionality design
- Conversation template definition
- LoFi and HiFi prototyping
- Moderating and analyzing user interviews
- Stakeholder management
Product vision
My ambition was to design a solution that would bring additional value to the customer by employing already existing solutions in a conversational and insightful manner where the user is able to explore and browse easily.
Allow users to understand their finances, receive specific insights and support, explore financing possibilities comfortable without the pressure of committing they might have in a 1:1 setting with a banking specialist.
Discovery and research
Analyse existing customer data to understand pain points in our solutions
Shadow customer service representatives to gather insights into their communication methods with customers and their daily tasks and pains.
Explore similar solutions to understand the market and what the makes a solution successful
Dive into data on most asked questions in the customer centre to uncover sensitive areas where customers need support.
During the project time line I have design and conducted 5 qualitative user test, from which 2 of these have included an eye-tracking portion.
Vision and concept design
When coming to this, it is a big task that can feel overwhelming. My strategy was setting some must-know topics and start defining those:
- where should it be placed: when does the user need it?
- what should it answer: what are the users calling about and why are they calling?
- how should it answer and what should the user leave with
Thes topics helped me kickstart the process and outline the vision, and together with the solution design drafts and exploration, I had a tangible concept and vision.
Exploratory LoFi visuals to showcase contextual placement and proactivity functionality
As the project moved forward and we conducted more user interviews and gained usage insights as well, the vision was revisited, compared against data and refined.
I ended up following a similar approach, focusing on
- advanced conversational capabilities
- user intent recognition and analsysis
- customer data based insights in request
- integrated calculation tools
Solution design
My approached focused on first drafting out the initial experience of key areas. To achieve this i mapped a customer journey and highlighted potential pain points. These pain points were later validated with moderated qualitative user interviews.
High level customer journey example with highlighted pains
My process focused on drafting multiple such journeys within the AI solution, and then establish pattern.
The outcomes were:
- a structured conversation template for content creators to follow and be aware of in their work
- an initial interface and needed components
- initial prototypes to conduct user testing on
Solution design: Conversational capabilities to clarify and troubleshoot
Data shows, users meet the AI solution as a regular chat, chatting in their natural style or in very structured ways. I have encountered use cases where users are very direct and use very systemic messages, almost key words, or others who tend to send very descriptive and detailed explanations. How can we design a solution that will cater for all?
Solution:
- asses the user’s intent
- determine if the intent implies multiple option
- deliver questions to support the customer in proving the relevant details
Solution design: Contextual suggestions in the solution depending on the origin screen
A highlight in my research has been that despite users knowing how to solve certain tasks, they still prefer to contact customer service for reassurance.
Solution:
- populate the knowledge base with information users generally feel unsure about to cater for the further needed explanations
- set the right expectations for the user by suggesting what the solution can support with
- meet them where they are in their task by matching their context.
Solution design: Guide the user step by step in relevant scenarios
The user has asked for guidance with a specific task, they might need to revisit the instructions during their task, so we need to make it convenient and digestible.
Solution:
- clearly indicate the conversation is running and the assistant is active
- keep track of the conversation by showing a preview of the last sent message
- meet the users pace by delivering instructions one by one and sync the steps to the screen the user is on
Conclusion
The project is still running and design on it is continuous. I am currently working towards fulfilling the vision. For this, i am aligning with journey design teams to explore feasibility and needs for enhanced functionality.
You can find the MVP version live in your Danske Bank mobile bank.
