Create Your First Project
Start adding your projects to your portfolio. Click on "Manage Projects" to get started
Scotiabank Chatbot
Designing a conversational assistant that scales trust
Context
In 2022, chat support was the #1 requested feature by Scotiabank's mobile customers. The business needed a chatbot that could reduce contact centre load, improve resolution speed, and modernize digital service—without eroding user trust.
Challenge
We were working with an AIML-based assistant, not an LLM—so flexibility was limited. Our flows covered sensitive, high-volume issues like password resets and credit card disputes. We had to align legal, compliance, and accessibility teams without compromising the user experience. One stakeholder pushed for overly flattering language, which conflicted with our voice principles.
My Role
As lead content designer, I was responsible for the assistant's overall content strategy, voice, and experience quality. That included writing and testing dozens of conversation flows, defining tone and interaction standards for scalable chatbot writing, leading cross-functional alignment with legal, compliance, accessibility, and brand, and documenting guidelines and flows in Confluence to support future reuse.
Key Moments of Influence
I ensured brand voice integrity by pushing back on stakeholder requests for overly flattering language—reinforcing that competence builds trust more than praise. I aligned legal and compliance on tone by prototyping sample flows and running review workshops—turning approvals into a partnership instead of a bottleneck.
Strategy
We framed the assistant as a service layer, not a character. That meant prioritizing predictable, scannable patterns over clever dialogue, designing micro-interactions that clarified next steps and minimized dead ends, creating a tone guide that balanced approachability with accountability—especially in edge cases, and making accessibility a design input, not a QA checkbox.
Execution
I authored 25+ flows across login issues, payments, card management, and profile updates, built a reusable conversation system for openings, confirmations, and closings, partnered with data team to test and tune intents within the AIML logic framework, led accessibility reviews with screen reader users, and created tone and flow documentation to support future automation efforts.
Impact
85% of user inquiries resolved without human escalation, 30% reduction in contact centre volume, framework reused across other assistant products internally, and voice, structure, and clarity consistently praised in internal UX reviews. Recognized with a 2023 Digital Transformation Award by IT World Canada for modernizing Scotiabank's customer experience.
Why it matters
We didn't try to make the assistant fun. We made it reliable. That clarity—plus structure and ethical tone decisions—built something users could trust, and internal teams could scale.
Date
2022
Project type
Conversational UX & content design
Role
Lead content designer
Location
Toronto
Problem: Improving automated customer interactions with better UX writing.
Process: Designed chatbot responses with clarity, accessibility, and engagement in mind.
Outcome: Increased engagement and efficiency, recognized by the 2023 Digital Transformation Award.