Making a banking platform legible to the people building it

Digital Banker aimed to unite fragmented banker tools. Bankers needed less searching, re-keying and chasing; customers needed fewer pauses while the bank reconstructed information it already held.

Tool-by-tool migration could hide broken handoffs behind a cleaner interface. Each squad could optimise its surface while the service between squads remained nobody’s problem, preserving the effort the program intended to remove.

The work crossed five views: customer, banker, operations, policy and data, and technology. Each used different boundaries, language and meanings of ‘complete’: submitted, checked, approved or visible in another system.

Evidence was distributed across bankers, operational specialists and technical teams. We also had to distinguish protective controls from workarounds caused by incomplete information or unclear ownership; simplification could otherwise weaken a control or move effort out of sight.

Consolidation had meant moving the most-used tools into one place first. That made migration volume visible but left service quality implicit.

I mapped one lending journey across five lanes and traced information through seven handoffs in the recreated structure. Three re-keying points concentrated follow-up, uncertainty and avoidable checking.

The question changed from ‘Which tool moves first?’ to ‘Which service break disappears first?’ Program leadership prioritised the three handoffs before broader interface consolidation, giving product, architecture and operations a shared order without pretending every dependency was settled.

I organised discovery around unknowns, interviewing bankers and specialists and walking live work with synthetic examples. Each inferred handoff was checked with its sender and receiver, then marked observed, reported or assumed.

I layered the blueprint from people and work to policy, data, systems and evidence strength. We modelled seven handoffs becoming three by carrying verified context forward, then proposed ‘guide, hand off, close the loop’ for movement between Digital Banker and the origination system; it has not been implemented uniformly.

I did not design a complete target interface, blueprint every lending variation or score teams against the map. These would create false certainty before policy, data access and architecture decisions were settled, or turn dependency mapping into blame.

People had to recognise their service in the blueprint. Lane-by-lane playbacks let bankers, operations specialists, policy partners and engineers correct details before product, operations, policy, data and technology leads named what they received, trusted, changed and owned at each break.

The leadership session sought a sequencing decision, not agreement. Leaders chose between two orders and prioritised shared handoffs; later, squads used the blueprint in briefs and dependency conversations without my facilitation, an observed adoption signal, though consistency across all eight squads was not measured.

Tool migration was recast around three cross-squad handoffs, with named product, operations and technology participation. Verified context travelling with work and one visible owner for exceptions became committed direction; reduced re-keying remained projected while delivery continued.

Customers were expected to repeat less context and face fewer unexplained pauses, but this was not measured. Designers also considered customer, banker and operational consequences in briefs and critiques; this was observed, not measured across all eight squads.

I now make a whole-service map the first deliverable in consolidation. Showing every handoff turns sequencing from negotiation between tool owners into a decision about which break to remove first; an evidence register beside it keeps every line challengeable as policy and architecture change.