Customers Bancorp, Inc.CUBB

Customers Bancorp has deployed 600+ AI agents, saved 46k hours, and licensed 100% of staff.

“6 © 2026 C USTO M ERS BANC O RP, INC . / ALL RIG HTS RESERVED Agentic Integration Approach AI: Experiencing Transformational Change Across the Bank 1. Represents target-state operating goals; initiatives are in progress and not realized 2. Hours saved reflect self-reported estimates from survey of employees using AI tools. Survey data as of July 3, 2026 3. Includes custom GPTs and other similar AI agent products across major platforms used in the last 6 months 4. Includes FTEs mobilized across AI transformation initiatives Increase Revenue Reduce Risk Improve Productivity Top-Down Strategic Transformation Priorities1 Loans 7-day close, down from 30+ days Deposits Commercial onboarding in minutes, not hours Payments cubiX agentic interface enabling the processing of nearly 2 billion client API requests by YE’2026 Front Office Operations RiskFinance & Accounting Compliance Data & Insights 46,000+ Hours saved2 600+ Agents deployed3 100% AI-licensed 40 AI mobilization team4 Bottom-Up Use Cases Key Results Select Partners Cycle seeks to weave AI into the Bank’s operating fabric Pair engineers + bankers = agentic pods Understand shadow the real workflow Prioritize highest- impact opportunities Build agents on our own data & systems Validate prove it, measure impact Absorb ship into the operating fabric Agentic workflow”

Customers Bancorp deploys AI across operations with 500+ agents built and hours saved.

“6 © 2026 C USTO M ERS BANC O RP, INC . / ALL RIG HTS RESERVED AI: Operationalizing Across Customers Bank ✓ Goal: be a leader in AI adoption among regional banks ✓ Measured in outcomes: cycle time, cost per transaction, error rates ✓ Small enough to move fast, large enough to invest with intent ✓ AI is the most significant opportunity in a generation for a bank of our size Increase Revenue Reduce Risk Improve Productivity AI in Production Strategic Vision Efficiency Augmented by AI Early Results $910 $981 2024 2025 Revenue per Employee1 $ thousands $286 $346 2024 2025 Pre-tax Profit per Employee2 $ thousands $27 $28 2024 2025 Avg. Assets per Employee3 $ millions Of team members AI licensed 75% Agents4 built by team members 500+ Hours saved5 28,000+ 1. Revenue / employee = (Net Interest Income + Non-Interest Income) ÷ Average FTE. Average FTE is the simple average of the four quarter-end FTE counts 2. Pre-Tax Income / employee = GAAP Income Before Income Tax Expense ÷ Average FTE. Average FTE is the simple average of four quarter-end FTE counts 3. Average Assets / employee = Average Total Assets ÷ Average FTE. Average FTE is the simple average of four quarter-end FTE counts 4. Includes custom GPTs and other similar AI agent products across major platforms used in the last 6 months 5. Hours saved reflect self-reported estimates from survey of employees using AI tools FOUNDATION Governance & Data 2024 2025 2026 2027+ PRODUCTION Piloting Use Cases PHASE 1 Deploying Use Cases and A”

Customers Bancorp has AI in production for code writing, automation, underwriting, and risk monitoring while training employees to build agents.

“We are starting to move past pilots. Across the bank, AI is in production today — writing code, automating manual processes, drafting and routing customer communications, accelerating underwriting and operational reviews. This is not AI theater. We measure it in cycle time, cost per transaction, and error rates, not just in adoption statistics or training. And we are not waiting for a central technology team to build everything: we are training every employee to build their own tools, because the bank that wins the next decade will be the one whose people learn to direct AI well — builders and managers of agents. Most banks will compete on productivity. We intend to excel there — but we intend to differentiate ourselves on the other two: Increase Revenue. AI is the lever that lets us transform the customer experience at scale, protecting and extending our industry-leading NPS of 81, opening new revenue streams inside cubiX and our specialty verticals, and standing up new business initiatives with an AI-first, high operating leverage mindset from day one rather than building cost structures we later have to unwind. Reduce Risk. We are embedding AI tools across the first, second, and third lines of defense, with the goal of moving from sample-based to continuous monitoring — which will allow us to audit what we can audit without resource limitations rather than only what we should audit. Done right, this turns risk management from a cost center into a moat around our business m”

Customers Bancorp has AI in production for code writing, automation, communications, underwriting, and risk monitoring.

“We are starting to move past pilots. Across the bank, AI is in production today — writing code, automating manual processes, drafting and routing customer communications, accelerating underwriting and operational reviews. This is not AI theater. We measure it in cycle time, cost per transaction, and error rates, not just in adoption statistics or training. And we are not waiting for a central technology team to build everything: we are training every employee to build their own tools, because the bank that wins the next decade will be the one whose people learn to direct AI well — builders and managers of agents. Most banks will compete on productivity. We intend to excel there — but we intend to differentiate ourselves on the other two: Increase Revenue. AI is the lever that lets us transform the customer experience at scale, protecting and extending our industry-leading NPS of 81, opening new revenue streams inside cubiX and our specialty verticals, and standing up new business initiatives with an AI-first, high operating leverage mindset from day one rather than building cost structures we later have to unwind. Reduce Risk. We are embedding AI tools across the first, second, and third lines of defense, with the goal of moving from sample-based to continuous monitoring — which will allow us to audit what we can audit without resource limitations rather than only what we should audit. Done right, this turns risk management from a cost center into a moat around our business m”