Building stronger foundations for AI in financial services
In 2024, much of the focus was on potential—new use cases, generative models and early experiments—while this year, the tone was more pragmatic. Banks are recognizing that success isn’t about how many pilots they launch, but how effectively they can govern them.
There’s no single roadmap for AI adoption. Each organization’s data, infrastructure and strategy defines its own path. What’s becoming universal, however, is the recognition that good governance drives long-term value.
For most banks, AI priorities have shifted from moving quickly to building the right foundations for scale. Those seeing the strongest results today are the ones that invested early in data governance, AI risk management and AI regulatory compliance. Many banks acknowledged that their early proofs of concept delivered uneven results when tested on real data, a reminder that success depends as much on preparation as on innovation.
Creating better AI governance
The most discussed topic at MoneyLIVE 2025 was the challenge of building reliable AI governance frameworks.
Across several sessions, speakers emphasized that sustainable innovation requires structure: clear accountability, consistent oversight and transparent measurement of results.
For many banks, this begins with infrastructure updates like modernizing data pipelines, upgrading architectures and ensuring data quality. In several discussions, leaders admitted they are still wrestling with legacy systems that make even basic data streaming and API integration difficult, slowing progress toward scalable AI. Without these basics, even the best AI models fail to perform consistently. To move faster with AI, companies must first establish control.
Governance isn’t a constraint, it’s what enables scale. When institutions establish the right controls early, they can industrialize AI safely and sustain innovation over time.
Tackling agent sprawl and AI oversight
Off stage, hallway discussions reinforced the same themes.
CIOs and CTOs spoke openly about the growing complexity of AI operations: too many proofs of concept, too many overlapping tools and too little oversight.
The new challenge is agent sprawl, as teams across the business experiment with autonomous systems without a unified governance model. In many organizations, these agents are being created faster than they can be monitored. Several attendees described internal teams launching dozens of small automations and “agents” without shared oversight, raising new concerns about cost, data usage and duplication of effort.
Banks are now looking to regain oversight by defining clear ownership, improving data controls and setting limits on usage and cost. Through AI operating models like GFT’s AgentTech Governance Framework, institutions are beginning to bring structure back to innovation.
Success depends less on chasing the next big use case and more on strengthening the systems, data and culture that make those use cases sustainable.
Where banks should focus their AI efforts in 2026
Looking ahead to 2026, financial institutions are entering a new phase of digital transformation in banking, one defined by discipline, measurement and collaboration.
Leaders are learning that it’s better to do fewer things well than to spread resources too thin. Many are re-evaluating their AI budgets and roadmaps, focusing on fewer, higher-value projects and delaying others until their governance models are mature. The institutions that start slow and focus on strong foundations today will be the ones that accelerate tomorrow.
AI’s next chapter will be written by organizations that approach innovation with purpose and control. With deep experience in AI governance frameworks, data governance and large-scale delivery, GFT continues to help banks move confidently from pilots to production.
If your institution is preparing for what comes next, get in touch to learn how GFT can help turn AI strategy into action with clarity and confidence.
