Bank FOMO drives AI spend past $40 billion: Accenture

Accenture's findings come as banks continue to pour money into AI, with spending exceeding $40 billion last year by one estimate. Despite this, only 20% of bank leaders report widespread, sustained value from AI initiatives, pointing to scale challenges. Mike Abbott, Accenture's global banking lead, attributes the spending surge to a fear of missing out, noting that while confidence in AI's potential is high, measurable returns have yet to materialize.

The survey, which gathered responses from 212 retail banking and 110 capital markets C-suite executives across 20 countries between April and June, highlights a gap between investment and outcomes. Abbott explains that task-wise productivity gains of 10–15% per person don't translate to system-wide productivity, as most banks haven't reconfigured work around AI. Banks are currently focused on expense benefits rather than revenue generation, finding value in call center operations, underwriting, marketing, and regulatory reporting automation. However, Abbott believes this will change as banks learn to optimize models, using smaller open-source models for document ingestion and building their own transactional foundational models.

The stakes hinge on whether banks can move from individual task improvements to rethinking entire processes. Abbott expects ROI to take center stage next year, with a shift from serial to parallel thinking, where AI agents can handle multiple steps simultaneously in areas like mortgage processing and software development. If a few elite banks demonstrate mainstream success, others are likely to follow, given the hyper-competitive nature of banking. The key test is whether banks can invest in their people and avoid consensus-driven paralysis to achieve system-wide productivity gains.