Wells Fargo Innovation Group · Enterprise innovation · Virtual Assistant + R&D mechanism
Trust, intent, and escalation: the enduring questions behind AI Product strategy
Wells Fargo Innovation Group was the bank's enterprise innovation organization, focused on emerging technology, digital customer Experience, and new business models. The Virtual Assistant was one of the Product and Experience initiatives I worked on within the group; I also created its R&D intake and output mechanism.
Historical enterprise innovation / AI Product chapter
90-second case
- Problem
- Forward-looking enterprise work needed to connect emerging-technology and Experience exploration to the scale and constraints of the core bank. Within that mandate, a financial assistant had to support meaningful customer goals within the technology and institutional-trust constraints of its period.
- Mandate
- Within Wells Fargo Innovation Group, serve as Lead UX Designer for The Virtual Assistant and create a playbook and process for the group's R&D intake and output queue.
- Decision
- Give innovation work an explicit receiving path; for the assistant, frame the Product around customer goals, financial health, relationship, and escalation—not chat alone.
- What happened
- The work produced an Innovation Group R&D mechanism and a Virtual Assistant direction spanning research inputs, a three-dimensional framework, personas, vignettes, two prototype directions, and omnichannel intent.
- Why it matters now
- It shows enterprise innovation and AI Product/Experience inquiry before the current generative-AI cycle.
My role
How I contributed.
I led Experience direction for The Virtual Assistant within Wells Fargo Innovation Group and designed the group’s R&D intake and output process.
I worked with Product Strategy, Product Management, Design, Engineering, Data Integration, Technology, and innovation stakeholders to move emerging ideas toward useful decisions.
Current reconstruction
The core mechanism.
A present-day visual of the system described in the case.
Judgment under constraint
Decisions that shaped the work.
Decision
Define the relationship and customer goal before the assistant surface.
- Why it was hard
- The period's familiar interaction patterns encouraged a narrow automation or conversational frame, while a financial relationship carries trust, context, and consequential goals.
- My judgment
- Use a three-dimensional model spanning Financial Health, Financial Savvy, and Relationship with the VA, then connect personas and scenarios to proof-of-concept priorities.
- Consequence
- The work narrowed into two initial journeys and prototype directions inside a channel-agnostic framework.
Major initiative · Wells Fargo Innovation Group
The Virtual Assistant was framed as an omnichannel relationship, not an isolated chatbot.
The original goal was to define a framework and platform for an ideal omnichannel virtual assistant that challenged assumptions about how machines could enhance human experience. The challenge explicitly considered the technology available and familiar at the time, existing AI behavior and human interaction, and perceptions of large banks and existing relationships.
Research and ideation included interview questions, speculative film and technology examples, a personality survey, popup-lab analysis, and secondary research. The team established common terminology, modeled relationships with existing and future AIs, and used personas, persona-to-VA relationship models, storyboarded vignettes, and user feedback to prioritize proof-of-concept stories.
- 01Customer goal
Financial-health aspiration · situation · relationship
- 02Relationship model
Financial Health · Financial Savvy · Relationship with the VA
- 03Scenario
Persona · context · storyboarded vignette
- 04Prototype direction
Visual / Interactive · Conversational with a View
- 05Omnichannel frame
Mobile plus voice · channel-agnostic system
Experience strategy
“Infinite Space” treated conversation as one mode inside a broader work surface.
The historical UX point of view envisioned multiple coexisting conversations and interactions, saved interactions, shifting contexts, and an assistant that could always show the user something. The goal moved from mobile-first toward mobile plus voice inside a channel-agnostic framework.
That period concept is relevant because it resisted a chat-only destination. It was a product vision from an earlier technology era, before modern LLMs and agent systems.
Broader Innovation R&D portfolio
The assistant was part of a broader system for shaping emerging opportunities.
My Wells Fargo role extended beyond one intelligent-assistant concept into emerging Product strategy, business and Experience definition, prototyping, reusable innovation systems, and receiving-team decisions.
The portfolio below is grouped by leadership purpose rather than technology so the repeated operating pattern remains visible.
Emerging Product concepts
The Flow Project · VR financial health
I pitched and led Product strategy and design for an immersive financial-health concept intended to make money flows understandable and surface personalized actions, partnering across Product, content, 3D Engineering, animation, and strategic relationships. The work remained an evolving concept and prototype direction.
Visual Data Aggregator · AR
I pitched and shaped an augmented-reality financial-data visualization concept connecting transaction flows to useful stories and business contexts through Product and business strategy, stakeholder alignment, partnerships, and Product Design oversight.
Future branch prototype
I pitched a future-branch concept to test whether physical locations could become differentiated customer environments rather than simply close, combining strategic research, concept development, stakeholder persuasion, Distribution Strategy partnership, and initial prototype planning.
Systems for scaling innovation
Kainos Design System
I spearheaded a reusable Innovation Group system spanning principles, interaction and component patterns, typography, grids, visual structure, design assets, and a longer-term direction for Design-to-Engineering integration across branded and unbranded R&D concepts.
Digital Product Design Playbook
I created a repeatable Design Thinking and Lean UX guide for Product Managers and Designers, connecting problem framing, experimentation, tools, templates, and cross-functional Product and Design work.
Innovation Design Critiques
I established a recurring critique mechanism for approximately 10–15 designers to improve R&D quality, collaboration, and mentoring.
Business-line experiments
Semantic graph / FIBO / ML proof of concept
Explored how semantic financial data and machine learning could turn a technology premise into a clearer business and Product recommendation.
Eastdil Secured innovation workshop
Used a structured innovation workshop to move from a business problem through discovery and concept development toward a receiving-team decision.
Wells Fargo Advisory concept
Shaped a referral and advisory concept through customer and business framing, Product definition, and a recommendation for how the opportunity could move forward.
What I now see in it · retrospective
Context, authority, visibility, and human continuity remain the hard Product questions.
Today's systems are far more capable, which makes the original Experience questions more consequential. A system that can interpret, propose, and act needs explicit context, permission, consequence, confirmation, escalation, and recovery—not simply more fluent conversation.
The earlier inquiry now informs how I think about agentic Products: context, permission, consequence, confirmation, escalation, and recovery must be designed together.