Branch traffic has declined as more consumers adopt digital self-service. Yet even with fewer in-person visits, customers still rely on their local branch for complex transactions, issue resolution, financial advice and cash services.
Many leading banks continue to view branches as essential for growth and customer experience. In July of 2025, JPMorgan Chase & Co. opened its 1,000th branch since launching its 2018 expansion plan. PNC Financial Services Group announced plans to double its new branch openings to nearly 200 across six states. Bank of America Corp. expects to open more than 150 financial centers by the end of 2027.
As the role of the branch evolves, so do the strategies banks use in their teller lines. Simultaneously, institutions are facing tougher operational challenges, particularly around fraud.
Reducing Fraud Risk at the Teller Window
More than half of all deposit fraud occurs at the teller window, where human judgment carries the most weight. For bank leaders, this represents a growing operational risk.
Fraud prevention should begin at the point of disbursement and continue through the entire clearing process. Automated verification and detection tools reduce manual work, lower false positives and support tellers in making informed decisions.
Research shows that the human brain can process only about four chunks of information at once, and fewer as complexity increases. Machine learning (ML) models, by contrast, analyze thousands or even millions of factors simultaneously. For example, signature and check-stock image analysis involves millions of pixels, with each being a data point.
A real-time fraud detection system instantly evaluates over the counter (OTC) transactions, including small dollar items fraudsters use to test vulnerabilities. These systems analyze account history, transaction trends, prior items and location data to assess risk before completion.
When no issues are detected, tellers can proceed confidently. If something appears suspicious, the system flags it for escalation which allows branches to move quickly without slowing service.
Banks can further tighten their defenses by implementing account and check validation tools that support image-based verification and authentication. Machine learning and advanced algorithms help institutions continually improve by identifying gaps, surfacing low-confidence or uncertain data and learning from new patterns.
Machine learning allows systems to identify gaps in their own understanding and provide feedback when they encounter low-confidence or uncertain data. This feedback strengthens the model, turning previously uncertain patterns into recognized, actionable signals.
Key structural pillars to maximize machine learning effectiveness include:
1. Mixing multiple machine learning methods. Combining approaches improves pattern recognition and predictive accuracy.
2. Deploying sensory monitors. These flag outliers, threshold breaches and unusual events early.
3. Expanding the data context. Including broader and more data improves precision.
4. Combining transaction and image analysis. Joint scoring models create a more complete risk picture.
5. Streamlining workflows. Efficient electronic review and resolution speed alert handling and reduce errors.
By applying these pillars, institutions shift from reactive to proactive fraud detection and continuously improve defense strategy.
Old Tools, New Tactics
According to the Association for Financial Professionals (AFP), 63% of organizations experienced attempted or successful check fraud last year. Fraudsters target checks because they’re easy to manipulate and are often tied to large recurring payments.
Some criminals wash stolen checks to remove ink and alter payees or amounts. Others cook checks by digitally editing images to create multiple fraudulent versions.
These items are frequently written for small amounts to evade detection and prolong the scheme. This repetition makes it difficult for banks, especially those relying on legacy systems, to spot fraud early enough to stop losses.
Beyond check fraud, transaction fraud remains a significant challenge. It includes deposits, return deposit items (RDIs), remote deposit capture (RDC) and duplicate ATM deposits targeting new, dormant or closed accounts.
Staying Ahead of Fraud Protects Banks and Customers
Balancing digital transformation with an evolving branch strategy is challenging, particularly when fraud pressures continue to rise. By leveraging automation, financial institutions can strengthen protections at both the physical and digital front lines, ultimately safeguarding assets, preserving trust and ensuring customers feel valued and secure.
Previous