Heidi Custers, Strategy and Transformation Director at Backbase, sits down with Wessel Oosthuizen, co-lead of the AWS AI and Data team for Deloitte Africa, to work through why bank AI programs stall - and what it actually takes to fix them.
Their starting point is a Deloitte statistic from a recent report that frames the whole conversation: 84% of companies haven't redesigned a single job around AI. Wessel argues that banks are making real technology investments while leaving their people and operating models exactly where they found them. As a result, the tech moves, but the organization doesn't.
Heidi pushes on the practical reality for regulated institutions - how to move fast without gutting compliance, why standalone AI pilots rarely stick, and how a front-to-back wedge approach lets business unit leaders reimagine their own processes without waiting for a top-down mandate.
They also get into wealth management, regional AI maturity, and the limits of Western-trained models, arguing that the real constraint on AI in banking isn't the technology; it's whether the people, the processes, and the data behind it were built for the world the bank operates in today.
Step into the future of banking -> Check out our content hub
Timestamps
00:00 Introduction
01:12 The Deloitte 2026 State of AI report - what 84% tells us
04:49 Employees using AI in secret - the hidden adoption problem
06:24 Why standalone AI pilots fail to move the needle
09:12 Building a fast operating model inside a regulated bank
14:05 The wealth management use case - context graphs in practice
18:31Starting small - capturing advisor knowledge before scaling
20:00 Regional differences - innovation fatigue in Europe vs. Africa's leapfrog moment
25:38 Why AI models trained on Western data get Africa wrong
29:09 The next decade belongs to better operating models, not better channels
