← All work01 · BANKING · OPS

Performance culture and AI tooling delivered up to 40% productivity uplift in a 1,200-person operations department

Client
Leading German regional bank
Year
2024
Role
Engagement Manager. Owned the AI productivity workstream end-to-end. Co-led the performance culture rollout with the client COO and reported into the program steering committee.

Context

Post-pandemic cost-to-income pressure had stalled a long-running efficiency program. The bank had invested in digital tooling and early GenAI experiments but couldn't point to measurable throughput gains. Adoption was uneven, performance management inconsistent across cluster leads, and the productivity ceiling looked stuck despite the investment.

Problem

Leadership needed a measurable uplift inside a fixed program horizon — and the operating muscle to keep delivering after the consultants left. Three things were missing: a shared performance baseline, a coaching cadence at team-lead level, and a focused AI rollout with real adoption mechanics behind it.

Approach

  1. 01

    Diagnostic

    Shadowed teams across the department, ran ~40 interviews, segmented operations by process type and uplift potential. Identified eight high-leverage process clusters and ranked them by speed-to-impact.

  2. 02

    Performance culture

    Installed a shared management system across all clusters — common KPIs, daily 15-minute huddles, transparent dashboards visible to every team member. Trained ~60 team leads on coaching rather than supervision. The unlock wasn't the metrics; it was making team leads owners of their numbers.

  3. 03

    AI-enabled productivity layer

    Selected and rolled out a focused stack: GenAI for document drafting and summarization, intelligent document processing for inbound mail, knowledge retrieval for the service desk. Piloted in three clusters with pre/post baselines and a control group. Built an ambassador network of ~25 power users to drive peer-to-peer adoption — the only mechanism that consistently sticks.

40%

peak productivity uplift in pilot clusters

8

high-leverage process clusters identified

Outcome

Stack

Power BIMS TeamsExcelMiroConfluenceMicrosoft CopilotChatGPT

Next case10From scattered pilots to a board-approved AI roadmap at a leading Landesbank →