Safe Haven Internet Banking System Redesign

By 2024, the platform had accumulated a paradox common to fast-growing fintechs: the underlying infrastructure was genuinely excellent, but users couldn't access what was under the hood.

In Q4 2024, the decision was made at the executive level: a complete redesign of Safe Haven's entire Internet Banking System.

I led a comprehensive review of the existing Internet Banking System across every touchpoint, covering both individual and business surfaces. This was a deep, structured team investigation across three dimensions: design audit, UX audit, and performance audit.

The findings were worse than expected. The product had accumulated years of additive feature development without a coherent design system to govern it.

The audit gave us the internal diagnosis. We validated it externally through support ticket analysis (a systematic review of complaint patterns) and user testing (structured observation of how real users navigated the existing system).

With a diagnostic this clear, the risk was moving too fast. Audit data can drive design into over-correction: fixing every finding, redesigning everything at once, optimising for the audit rather than for the user. The strategy we set was deliberate: use the findings to set direction, but let human outcomes anchor every decision.

Before designing a single new screen, I made a foundational call: the design system comes first. Every piece of accumulated inconsistency across the platform could be traced back to the absence of a single source of truth. We were not going to repeat that mistake.

I built the design foundation from scratch, constrained by the existing brand but establishing proper structure where none had existed. The foundation covered colour, typography, spacing, icons, and elevation. I set the primitive layer personally, then delegated design token construction to the first designer and accessibility testing to the second.

Every component had to pass all validation criteria before being admitted to the library. No exceptions.

With the design system in place, redesign work began across all platforms. The process was deliberately non-linear: mobile and desktop ran in parallel, and insights from one platform informed decisions on another through continuous iteration and beta feedback cycles.

The mobile app was the primary touchpoint for individual customers and the most visible experience failure. The core focus areas were;

The web experience was designed in parallel with the mobile app, maintaining visual and functional consistency across platforms while accounting for the different interaction model and information density appropriate for desktop contexts. Desktop users have more screen real estate and a different mental model for navigation; the redesign respected this without diverging from the unified design system.

The most significant design challenge across all the platforms was the Corporate Banking compliance onboarding flow. This was a genuinely broken experience causing user drop-off and unsustainable support load.

The developer platform served technical teams and fintechs building on Safe Haven's BaaS infrastructure: users with high technical fluency and low tolerance for interface friction. The redesign addressed API access and key management, sandbox environment experience, and fintech onboarding, integrating the same compliance architecture built for business banking where applicable.

Beta launched in October 2025 with three participant groups: internal staff, selected external customers, and members of Safe Haven's community programme. The objective was not just to find bugs. It was to observe real behaviour in context and identify what the design had missed or underestimated.

Feedback channels were matched to user type. Business customers communicated through Slack, allowing for detailed, technical back-and-forth. Individual users provided feedback through in-app mechanisms and direct conversations. Internal staff surfaced edge cases that external users were less likely to flag early.

The beta-to-launch cycle produced meaningful, user-validated additions. Several features that had been planned but not fully specified were accelerated into the launch build based on signal from beta participants:

The results were unambiguous. Across every signal that had indicated failure before the redesign, the platform was operating at a fundamentally different level.

When I disclosed the full project scope to the team, the response was close to alarm. The scope was objectively large. What made the difference was not minimising it but contextualising it. I walked the team through the framework, the phased approach, how the design system foundation would multiply efficiency across all platforms, and the progress metrics that would make the work feel measurable rather than infinite. I got genuine buy-in, not reluctant compliance.

As Head of Design reporting directly to the CEO, I operated at the intersection of executive expectations and product reality. This meant making calls in the midst of authority and holding those calls when challenged.

The support ticket analysis and user testing were valuable and validated the audit findings clearly. But a structured, ongoing research practice established earlier would have surfaced the discoverability problem before features were built and failed. Research is most powerful when it prevents wrong decisions.