Wells Fargo Conversational AI

Company
Wells Fargo

Timeline
2018–2021

Images shown above were not my mockups, although they illustrate the final product.

From 2018 to 2021, I worked at Wells Fargo, designing product concepts for conversational AI, internal tools for conversation designers, conversation analytics and machine learning. All of these efforts served as the foundation for Wells Fargo's chatbot, Fargo, which launched in 2022.

Some of the challenges faced and overcame were working with emerging technology in a highly regulated environment where usage of off-the-shelf tools was often prohibited for security and compliance reasons. As a result, we designed and built many of the tools that were needed to build the Fargo chatbot.

Try out Fargo here... https://sites.wf.com/fargo/

Case Study

Business Context

Bank was interested in leveraging the potential of AI and automation
Third largest bank in the US
71 million customers using mobile banking

Constraints

Highly regulated environment
Trust affected by legacy accounting scandal
Major focus on risk management, which slowed the pace of innovation

Design Transformation

I performed competetive research and created the initial bot persona strategy

Benchmarking eloquence

Frequency-based word clouds from popular voice agents

Conversation writing

Championed Ethics and Values for AI
Led the design of voice-first multimodal prototypes for mortgages, with multi-disciplinary team (Google Assistant & Siri)

Multimodal mockup concepts

Multimodal prototype - Google Home & DialogFlow

Led design of internal conversation design tool
- Initiated cross-group “show and tell” huddles with conversation designers and PMs across three time zones, to share best practices and get regular feedback on usability and needs.
- Performed market research on conversational AI design tool interface trends
- Designed analytics views (both integrated with tool and external tableau dashboard), emoji support, simulator, various features...

Conversation tool design

Invision Studio prototype template

Led the design of transcript annotation for machine learning
- Reduced annotation time by 50% saving the company time and money
Created blueprint of the bank’s entire AI ecosystem with recommendations
- Blueprint and recommendations were used by product management to onboard executives and act on opportunities

Outcome & Data

Measurable savings of time and money
Increase in positive feedback on UX for conversation designers
Soft launch on Mobile and Web to 71 million customers
The tools and templates I designed were extensively used