Selected details withheld under NDA
Building evidence-based contact-centre personas at TD
NACO persona development, mobile-app CX reporting, and AI research
- My role
- HCD Research + CX Analytics Intern
- Timeline
- Jan 2025 - Apr 2025
- Study scope
- 10,000+ feedback records · four-person research team
On this page
Research challenge & role
This case study focuses on the NACO persona project, where the team lacked a shared, evidence-based understanding of its contact-center customers. Our goal was to turn existing feedback and call data into a usable customer model.
- January to early April 2025. I joined as the project began and helped deliver its first stage within four months.
- My role. I organized customer feedback, reviewed recorded calls, and contributed to evidence synthesis and persona construction with the project lead and two full-time researchers.
- Team contribution. The four-person team compared patterns across sources and delivered the first-stage personas. Their specific attributes remain confidential.
Understanding the evidence and users
Using Excel, I reviewed and organized more than 10,000 rows of customer feedback alongside hours of recorded-call review, secondary research, and frontline context.
- Why this evidence mix mattered. Comparing structured feedback against recorded calls, external research, prior persona work, and frontline context reduced the risk of defining personas from isolated metrics.
Synthesis & persona construction
We used FigJam to cluster evidence, compare recurring needs, behaviours, barriers, and contexts across sources, and turn the broad question into smaller synthesis targets. Differences that held across the evidence base supported four distinct personas.
- Validation boundary. These were first-stage personas. Further validation with frontline employees and representative customer scenarios remained an identified next step.
Other work at TD
- Mobile-app CX reporting. I co-developed biweekly reports in Qualtrics XM for a 20+ person mobile-app development group. Monitoring app-store ratings and synthesizing 50+ customer comments per cycle surfaced recurring experience themes for product discussions.
- AI and personal-finance discourse analysis. I independently analyzed Reddit discussions to identify patterns in terminology, trust, and perceived risk. I presented the findings as research and design implications for the design research team.
Final output & reflection
We delivered four evidence-based personas and presented the first-stage findings to the NACO team lead and team members. The artifacts remain a working resource for the NACO team.
- Delivered outcome. The personas gave NACO stakeholders a shared reference for discussing distinct customer needs and behaviours using the same evidence base.
- Lesson. Large datasets become useful only after the research question becomes smaller. Combining internal data, secondary research, prior work, and subject-matter expertise helped turn an overwhelming evidence base into a defensible output.
- With two more weeks. I would validate the personas with more frontline NACO employees, test them against representative customer scenarios, and create an action matrix connecting each persona to service opportunities, evidence gaps, and future research questions.