Your enterprise AI framework determines your returns

Q: Why are financial services firms leading in AI adoption but reporting poor ROI?

The root cause of underwhelming AI ROI in financial services is organizational rather than technological. AI has been layered onto processes designed before AI existed, and those processes were not built to capture what AI can do. Changing the outcome requires changing the framework, which requires a different kind of leadership commitment. Financial services has deployed more AI from pilot to production than any other industry, and has the most to gain from closing this gap.

Q: What are the three phases of the enterprise AI framework?

Phase one is the learning phase: AI sits alongside existing workflows, tools proliferate without standardization, and learning does not compound. Phase two is integration: leadership consolidates onto a platform, domain expertise gets encoded into the AI layer, and a flywheel builds. Phase three is redesign: processes are rebuilt with AI as a native element, teams get smaller and more strategic, and bottlenecks are addressed through iterative simplification.

Q: How does domain expertise become a compounding advantage in AI-enabled financial services?

Encoded institutional knowledge: underwriting rules, claims adjudication logic, compliance requirements — becomes a compounding advantage that competitors cannot replicate by purchasing the same foundation model. Every subsequent interaction becomes more informed than the last. For institutions working through AI governance financial services requirements, it is also what makes governed AI systems defensible over time.

Financial services leads every industry in AI adoption, yet most executives report underwhelming returns. Institutions adding AI onto unchanged workflows are unable to capture returns at scale.

Chun Schiros

Field CTO

AWS

What you'll take away from this session

Phase 1 learning builds literacy, not returns

Phase 1 of the AI framework is the exploration phase, where teams learn what works from pilots and institutional knowledge about AI capabilities accumulates. Without standardization across teams, AI sits alongside existing workflows and learnings don’t compound. Ultimately, the organization will have nothing to show for the investment except a collection of pilots that did not scale.

Platform consolidation is the second phase of the AI framework

Moving from experimentation to enterprise implementation requires leadership to choose a platform that lets shared investment compound across teams. The choice signals that AI is no longer an experiment, but how the business operates.

Metric visibility is a prerequisite for transformation

Organizations routinely claim high AI usage while their metrics tell a different story. Knowing where time is actually being saved and where bottlenecks remain is needed because the flywheel cannot be built without the signals providing the knowledge of what baseline to build from.

Phase 3 starts with one design question

The question is, “If we started with AI as a first-class citizen in the design, what would this process become?” The answer will almost always be structurally different, not just iteratively improved. When one part of a process becomes dramatically faster, the next constraint becomes visible, allowing teams to iterate over time.

Governance enables AI to go faster

In financial services, safeguarding trust is paramount. Governance is the mechanism that allows autonomous AI systems to operate without destroying that trust.

Financial services organizations have deployed AI from pilot to production more than any other industry. However, the same organizations who have deployed AI are also the most likely to report that returns have not matched expectations. Chun Schiros, Field CTO at AWS and former Chief Analytics Officer with over fifteen years in financial services, argues that this highlights challenges with the AI framework, not the technology.

Her framework maps three AI maturity phases:

AI sitting alongside unchanged workflows
AI embedded as a native element of how work gets done
Processes rebuilt with AI as a core design input

The transition from phase one to phase two is marked by encoding institutional knowledge such as underwriting rules, claims adjudication logic, and compliance requirements into the AI layer. That expertise becomes the compounding asset competitors cannot replicate by leveraging the same foundation model.

In phase three, processes are redesigned with AI as a native element. Smaller teams become more strategic, asking “if we designed this process from scratch today, what would it look like?” The answer will almost always be structurally different, not just incrementally improved. When one step in a process becomes dramatically faster, it immediately exposes blockers in the next step, allowing teams to iterate over time in a continuous pressure-and-release cycle.

FAQ

Why are financial services firms leading in AI adoption but reporting poor ROI?

What are the three phases of the enterprise AI framework?

How does domain expertise become a compounding advantage in AI-enabled financial services?

Transform the work that matters most

See how Domino helps the world’s most regulated enterprises build, scale, and govern AI-powered applications.

Watch demo Talk to an expert