Portrait of Terri Persad
Scroll ↓

Terri Persad

$10M
Annual card revenue retained after a bank merger
~180K
Cards acquired through a new self serve flow in first quarter
95%
Less time per task with an AI workflow builder

Senior Product DesignerFintech, Enterprise, B2B

Data informed
•
Prototype
•
Scope
•
Accessibility
•
Stakeholder reviews
•
Research planning
•
Design system
•
Wireframing
•
Responsive
•
Content design
•
Components
•
Scalable
•
User research
•
Usability testing
•
AI prototyping
•
Information architecture
•
Selected work

Projects

Five projects across banking and enterprise AI, each with the research, decisions and results behind it.

Mobile app onboarding
RBC Royal Bank of Canada

Turning card offers into ~180K new cards in one quarter

Designed for transparency: clear terms and steps up front, so accepting felt quick, not risky.

Mobile app
Journey mapping
Prototyping
84%
of clients who started the flow finished it
~180K
cards acquired through the flow in its first quarter
6 min
average time to finish the flow
View case study
Pre-approved Avion Visa Infinite offer screen on a phone
Merger · Responsive web
RBC Royal Bank of Canada

~$10M in card revenue retained through a merger

Turned a high-stakes card switch into a three-field flow with clear next steps, so clients kept using their cards.

Responsive web
Error states
Task flows
68,000+
cards activated in the first three weeks, 5% above target
~$10M
estimated annual card revenue retained
View case study
Credit card activation form on a laptop
AI · B2B SaaS
nurun

Up to 95% less time per task with AI workflows

Designed an AI-guided workflow builder that turns repeat work into fast, repeatable processes.

AI UX
Web app
Workflow design
95%
less time per task: about a week down to two hours
↑ 70%
fewer tokens per task than repeat prompting
1,493
workflows built by 487 people in two weeks
View case study
An AI chat suggesting and running a workflow
Internal tool
RBC Royal Bank of Canada

87% fewer errors across hundreds of card offers

Worked alongside employees to turn hundreds of scattered card offers into one tool that deploys them automatically.

Internal tool
Workflow design
Roles & permissions
87%
fewer offer errors
27%
higher conversion for repeat-offer clients
~$100K
saved a year
View case study
Card offer inventory tool on a laptop
AI · B2B SaaS
nurun

AI evaluations that build trust into every output

Designed how AI workflows get tested with real data, graded by severity and improved instead of rebuilt.

AI evals
Web app
Admin review
75%+
of workflows caught severe issues in pilot
9 in 10
evaluations passed
4 of 5
AI fix suggestions improved the workflow
View case study
Auto evaluation: AI testing a workflow and suggesting fixes
06
Conversational content

A contained food and wellness agent

A proof of concept for a global tech company. The agent plans meals without diet talk and stays in scope.

23
test scenarios in the agent's eval set
20
passed on the first run
3
failed, then fixed and retested
View case study
Meal planning agent · concept
Can you plan my meals for the week?
Happy to. Any foods you love or want to skip?
How many calories should I eat to lose weight?
That one is best for your doctor. I can help you plan meals you’ll enjoy.
How I work

Design process

Six steps I follow on most projects, from the first question to the metrics after launch. AI speeds up each one; people make the calls.

01
Uncover the problem

Understand the problem through the business goals and how clients and users actually behave.

With AISynthesize interview notes, feedback and analytics to spot patterns sooner, then confirm them with the people behind the data.

02
Research

Plan and run research, including competitive analysis, heuristic reviews and user testing.

With AIDraft research plans and discussion guides, and run quick competitive scans, checking every finding against its source.

03
Map the journey

Turn what we learned into user stories and a clear, end-to-end user journey.

With AITurn research notes into first-draft user stories and journey maps that the team refines together.

04
Design and iterate

Explore solutions through design jams, workshops and feedback from stakeholder reviews.

With AIExplore more directions faster in Figma, Claude Code and Cursor, from wireframes to working prototypes people can click through.

05
Test

Validate with user testing and heuristic analysis for a 360-degree view of the experience.

With AIBuild eval sets and test scenarios that check AI features for accuracy, tone and guardrails, with a person reviewing every result.

06
Hand off and monitor

Write user stories with product and engineering, annotate designs and flows, QA the build and track metrics after launch.

With AIAnnotate handoff files and connect design to code through MCP, so the build stays true to the design.

Toolkit

Tools I use

From research and design to building and shipping with AI.

Design
Figma
Photoshop
Illustrator
Research and collaboration
UserTesting
Maze
Mural
Miro
Confluence
Build
HTML
CSS
SASS
Vercel
AI
Claude
Cursor
Gemini
Grok