Share with your CIO
Bank of America is declaring the pilot era over. CTIO Hari Gopalkrishnan, speaking at the Semafor World Economy 2026 event, laid out the bank’s four-pillar AI shift: end-to-end process transformation (not task tweaks), enterprise-wide scale and reuse across roughly 3,000 internal processes, tighter governance, and ROI accountability before projects launch rather than after. With 30% of a $13.5 billion technology budget funding new initiatives, the bank’s AI strategy reorientation signals a deliberate move from experimentation to embedded, measurable business outcomes.
What this means for your business
The organizations most exposed to this story aren’t banks. They’re any enterprise that spent 2024 accumulating proof-of-concepts and is now defending a budget line that can’t show a return. Gopalkrishnan’s framing is a useful diagnostic: if your AI portfolio is still organized around discrete tools rather than end-to-end processes, you’re in the phase Bank of America just said it’s leaving. The question isn’t whether to make that shift, it’s whether your governance and FinOps infrastructure can support it when you do.
The scale-and-reuse principle deserves more attention than it typically gets. Most enterprises default to letting individual teams build their own AI applications, which produces fast early wins and chronic duplication. Bank of America is explicitly moving toward shared, enterprise-wide AI capabilities that any of its 3,000 processes can draw on, treating AI less like software procurement and more like infrastructure. That’s a meaningful architectural commitment, and it implies a centralized platform team, not just a center of excellence with advisory power.
Gopalkrishnan’s governance framing, “overdo it and you stall innovation, underdo it and you introduce risk,” sounds balanced, but it actually cuts against the instinct most large institutions have when AI scales: to route every deployment through a lengthy review process. The 44% internal mobility fill rate the bank attributes partly to upskilling is the underappreciated output here. If AI genuinely retrains your workforce rather than just automating tasks, your talent strategy and your AI strategy stop being separate conversations, and your CHRO becomes a stakeholder in decisions your CIO used to own alone.
Concept deep-dive: FinOps
FinOps, short for financial operations, is the practice of tracking and optimizing cloud and compute spending in real time rather than reconciling costs at the end of a budget cycle. Applied to AI, it answers the question Gopalkrishnan raised directly: how much does it cost to run a given model at scale, and what business outcome justifies that cost? Without it, AI infrastructure spending grows faster than the value it produces, which is the pattern most enterprises are quietly sitting inside right now.
Based on reporting from Bank of America tech chief shares AI strategy focus, originally published 2026-04-14 03:00:00.
