From AI Experimentation to Operational Impact
August 22, 2026 | Resources
ONSETTO EXECUTIVE PERSPECTIVE
August 2026
The Opportunity Is Bigger Than Automation
When bankers hear the term AI, many immediately think about automation. But the most valuable applications of AI are not about replacing people. They are about helping bankers spend more time banking.
- Accelerate commercial prospect research
- Create more personalized treasury and business banking proposals
- Improve customer onboarding experiences
- Reduce manual document review
- Streamline policy and procedure management
- Enhance operational consistency
- Improve call quality monitoring
- Identify growth opportunities across existing relationships
The common thread is simple: AI works best when it reduces administrative burden and allows employees to focus on judgment, relationship building, and customer service.
Three Common Mistakes Banks Make
- Waiting for Perfect Clarity
Many institutions delay adoption because they are waiting for complete regulatory certainty. Waiting too long can create uncontrolled experimentation rather than controlled innovation.
- Letting AI Adoption Happen Informally
Marketing, operations, and relationship management teams may already be using AI without established policies, approved tools, or governance standards.
- Focusing on Technology Instead of Workflows
The most common question is 'Which AI tool should we buy?' A better question is 'Which banking processes could benefit from AI assistance?'
The Maker-Checker-Verifier Framework
Every banking process contains three fundamental functions that bankers already understand:
- Maker
Creates the work: preparing a loan package, drafting a treasury proposal, creating a customer communication, or building an account onboarding plan.
- Checker
Reviews the work for accuracy, completeness, and compliance: identifying missing documentation, confirming policy requirements, verifying data accuracy, and flagging inconsistencies.
- Verifier
Provides final approval and accountability: approving customer communications, authorizing credit recommendations, and validating compliance reviews.
Four Levels of AI Adoption
| Level | Maker | Checker | Verifier | Banking Meaning |
|---|---|---|---|---|
| Level 0: Human-driven | Human | Human | Human | Work is manual and often inconsistent. |
| Level 1: AI-assisted checking | Human | AI | Human | AI helps find errors, gaps, or inconsistencies. |
| Level 2: AI-initiated, human-managed | AI | Human | Human | AI drafts or assembles the first version; humans review and approve. |
| Level 3: AI-human synergy | AI | Human | AI | AI creates and performs final systematic verification; humans focus on judgment-heavy review. |
Importantly, the goal is not to move every process to Level 3. The objective is to match AI capabilities with business value, customer impact, and risk tolerance.
Where Banks Should Start
The best AI initiatives share four characteristics: they are repetitive, document-intensive, measurable, and operationally important.
- Commercial prospecting
- Treasury proposal generation
- Account onboarding and switching
- Loan documentation review
- Policy management
- Call quality monitoring
- Vendor due diligence
- Board reporting support
- Customer communication drafting
Five Questions Every Bank Should Ask
- What business problem are we solving?
AI should address a clearly defined operational or customer challenge.
- How will this improve customer or employee outcomes?
The objective should be measurable value, not technology adoption.
- What role should AI play?
Should AI act as a Maker, Checker, or Verifier? Understanding its role helps define appropriate controls.
- What oversight is required?
Human accountability should remain clear throughout the process.
- How will we measure success?
Success should be measured through reduced cycle times, improved customer experience, lower error rates, faster onboarding, increased employee productivity, and stronger relationship growth.
Moving From Curiosity to Confidence
Banks do not need to become artificial intelligence experts overnight. They need a practical framework for evaluating opportunities, managing risk, and delivering measurable business outcomes.
The institutions that will succeed are not those deploying the most AI. They are the institutions deploying AI thoughtfully - improving workflows, reducing friction, strengthening customer relationships, and enabling employees to focus on higher-value work.
Because in banking, success is not measured by how much AI you implement. It is measured by how effectively you use it to serve customers, support employees, and grow relationships.