The challenge
The bank's B2B customer-onboarding process spans 126+ products, each with its own workflow. No single team could see the process end to end, so delays, hand-off failures and compliance gaps were reported after the fact — if they were reported at all.
Payment-exceptions handling had the same problem in a different shape: domestic and international transaction flows passing through both systems and human touch-points, with no objective record of where cases stalled.
How we went about it
- 1
Prove it on one product family
We connected to the workflow, core-banking and case-management systems, harmonised their event logs into a single case timeline, and delivered a working pilot on one product family within two weeks.
- 2
Scale to the full onboarding estate
With the event model proven, we extended it across every product: 1,600+ distinct activities and 50M+ events, discovered from the logs rather than drawn from interviews.
- 3
Extend to payment exceptions
The same approach was applied to payment-exceptions handling — 250+ activities across system and human touch-points, for both domestic and international workflows.
- 4
Turn findings into a roadmap
Bottlenecks, rework loops and compliance deviations were quantified by product and by team, and prioritised into an improvement roadmap the business could act on.
What we delivered
- Harmonised event log across workflow, core-banking and case-management systems.
- Discovered process models for every onboarding product and for payment exceptions.
- Bottleneck, rework and conformance analysis by product, team and transaction type.
- A prioritised improvement roadmap grounded in measured process performance.
Outcomes
What changed for the client
- An objective view of where onboarding actually stalls, by product and by team.
- Compliance and hand-off failures detected from the event log rather than reported after the fact.
- A single, evidence-based improvement roadmap instead of competing anecdotes.
- A repeatable method: the second process (payment exceptions) was onboarded far faster than the first.