Service Design at the Digital Wallet
Service Design · Product Discovery · Mixed Methods · CJM · Service Blueprint · Multilingual Research · Stakeholder Management · Fintech · B2C · Startup
Over 8 months (June 2025 to February 2026) I worked as Service Designer & Product Researcher at m10, Azerbaijan's leading digital wallet and part of Pasha Group's BirEcosystem. Comparable growth trajectory to Wise at Series B.
As the first service designer in the ecosystem, I built the practice through live product work. The work spanned the digital card, loans, cross-border transfers, loyalty programmes, and a few other areas. Much of it was coordination across product, analytics, marketing, growth, and CX teams, anchored to the user. Each project ran in parallel, and insights from one consistently seeded the next.
As the first service designer in the ecosystem, I built the practice through live product work. The work spanned the digital card, loans, cross-border transfers, loyalty programmes, and a few other areas. Much of it was coordination across product, analytics, marketing, growth, and CX teams, anchored to the user. Each project ran in parallel, and insights from one consistently seeded the next.
Project example
Digital card. End-to-end
CJM · Service Blueprint · In-Depth Interviews · Qualitative UT · Quantitative · Post-release CSAT
A product launching into a cash-heavy market with no loyalty bonuses, no physical card, and users who had little reference point for how a digital card works.
I ran the discovery end-to-end: first interviews, usability tests, and the management of Azerbaijani-speaking interviewers (I don't speak the language). I led analysis and synthesis, and held stakeholder alignment across design, product, analytics, and development.
At release the communications side sat with me too: push logic, activation comms, first-transaction conversion tracking, all inside the same service-design remit.
Result:
A product launching into a cash-heavy market with no loyalty bonuses, no physical card, and users who had little reference point for how a digital card works.
I ran the discovery end-to-end: first interviews, usability tests, and the management of Azerbaijani-speaking interviewers (I don't speak the language). I led analysis and synthesis, and held stakeholder alignment across design, product, analytics, and development.
At release the communications side sat with me too: push logic, activation comms, first-transaction conversion tracking, all inside the same service-design remit.
Result:
- 50% of active wallet users activated the card within 30 days.
- Around 30% of those completed at least one transaction. No loyalty bonuses, no physical card, in a market where digital payments were still unfamiliar to a large share of users.
Digital card. End-to-end
CJM · Service Blueprint · InDebth Interviews · Qualitative UT · Quantitative · Post-release CSAT
A product launching into a cash-heavy market with no loyalty bonuses, no physical card, and users who had little reference point for how a digital card works.
I ran the discovery end-to-end: first interviews, usability tests, and the management of Azerbaijani-speaking interviewers (I don't speak the language). I led analysis and synthesis, and held stakeholder alignment across design, product, analytics, and development.
At release the communications side sat with me too: push logic, activation comms, first-transaction conversion tracking, all inside the same service-design remit.
Result:
A product launching into a cash-heavy market with no loyalty bonuses, no physical card, and users who had little reference point for how a digital card works.
I ran the discovery end-to-end: first interviews, usability tests, and the management of Azerbaijani-speaking interviewers (I don't speak the language). I led analysis and synthesis, and held stakeholder alignment across design, product, analytics, and development.
At release the communications side sat with me too: push logic, activation comms, first-transaction conversion tracking, all inside the same service-design remit.
Result:
- 50% of active wallet users activated the card within 30 days.
- Around 30% of those completed at least one transaction. No loyalty bonuses, no physical card, in a market where digital payments were still unfamiliar to a large share of users.
What the card discovery seeded
Cross-product synthesis · Backlog seeding · Workshop facilitation
I routed card-research findings into several parallel workstreams: loans, transfers, loyalty, etc.
Loans discovery surfaced a behavioural blocker. Users hesitated on "Apply" because they read the click as a final commitment, when the actual agreement came several steps later. Card research had already mapped trust patterns in similar moments, so the insight landed without a separate needs-gathering cycle.
The loyalty workstream surfaced its own friction. m10 ran two loyalty systems in parallel: BirBonus, the ecosystem-wide programme, and lemons, the wallet's internal currency. Users didn't know when to use which. I mapped the confusion to specific user flows and recommended in-flow placement, which design then turned into a concept for where each loyalty signal should appear in the journey.
I routed card-research findings into several parallel workstreams: loans, transfers, loyalty, etc.
Loans discovery surfaced a behavioural blocker. Users hesitated on "Apply" because they read the click as a final commitment, when the actual agreement came several steps later. Card research had already mapped trust patterns in similar moments, so the insight landed without a separate needs-gathering cycle.
The loyalty workstream surfaced its own friction. m10 ran two loyalty systems in parallel: BirBonus, the ecosystem-wide programme, and lemons, the wallet's internal currency. Users didn't know when to use which. I mapped the confusion to specific user flows and recommended in-flow placement, which design then turned into a concept for where each loyalty signal should appear in the journey.
A startup moving fast doesn't need a research department. It needs someone who can get to the right question quickly, run the leanest study that answers it, and make sure the finding actually changes something
More cases