AI Customer Service Assistant

AI powered solutions for personalized assistance with the toughest financial problems.

THE CHALLENGE

Uncovered high call center costs and customer frustration due to jargon-heavy calls and limited agent availability, driving financial stress and operational inefficiencies. I aimed to design an AI assistant that simplifies support and reduces overhead.

MY APPROACH

RESEARCH

Analyzed 50+ calls and conducted user interviews, identifying information gaps and top agents traits for effective support. Created agent and customer user personas based on the aggregated research as well as the customer group the AI would be rolled out to. Customer journeys visualized how difficult current state was for customers and highlighted opportunities for the future process.

DESIGN

Brainstormed multiple AI personas inspired by top agents, mapping various conversation flows in Figma.
Prototyped and tested the personas and tones with stakeholders and agents through user testing and questionnaires.
Simplified validation for development by creating a quick glance persona visual to reference. Designed around the tight timeline, technical constraints, and risks of AI hallucinations. Unique rollout method allowed for collecting data from real customers but still midigating risk, allowing rapid iteration.

SOLUTION

Delivered an AI personality and tone that the customers respond positively to as well as a simple interface which can easily be interacted with any time of day.

BEFORE

AFTER

THE IMPACT

BUSINESS

Reduced call volume by 25%, cutting costs and boosting compliance by minimizing agent errors.

CUSTOMER

Avoiding hardship by interfacing with a trustful guide.

KEY WINS

Took skills I already had, adapting quickly to apply them to this new domain.

Established design standard with an emphasis on a scalable, human-first, and well documented process, setting the company up for success for future AI inititives.