Practical Applications of AI in Financial Institutions Part 2: Customer Service

Artificial intelligence is transforming customer service across financial institutions. In 2026, banks and credit unions are no longer asking whether to deploy AI in customer-facing operations. Instead, they’re actively working to implement it strategically to meet rising expectations for instant, personalized, 24/7 support. According to a Deloitte survey of U.S. banking customers conducted in January 2025, while 37% of respondents have never interacted with banking chatbots, those who have are driving rapid adoption, particularly among younger generations.

This quarterly series focuses on practical, real-world AI applications in financial services. In our first installment, we discussed AI use cases in lending. For this second installment, we’re examining customer service, a function where AI is delivering measurable impact while reshaping how financial institutions interact with their members and customers.

Where AI Is Making an Impact

1. Conversational AI: Chatbots & Virtual Assistants

AI-powered chatbots and virtual assistants have become part of the front lines for customer service for major financial institutions, handling millions of interactions daily and providing instant responses to routine inquiries.

Examples of Tools in Action:

• Bank of America’s Erica: Launched nearly a decade ago, Erica has become one of the most successful AI virtual assistants in banking. By August 2025, Erica surpassed 3 billion client interactions, with 20 million clients using the assistant to manage their finances. Erica provides personalized insights, proactive alerts, and real-time support directly within the mobile app and online banking platform, helping customers understand AI’s role in managing their finances.
• Wells Fargo: Wells Fargo’s generative AI assistant, powered by Google Cloud, has demonstrated remarkable security and scale. By April 2025, Fargo handled 245.4 million interactions with zero personally identifiable information (PII) leaks. Fargo handles customer queries without requiring handoffs to human agents for routine matters, focusing on balance inquiries, transaction history, and basic account management.
• JPMorgan Chase’s AI Platform: JPMorgan Chase has deployed over 30,000 employee-built AI assistants, with its internal AI platform now used by 1 in 2 employees daily. The bank has implemented over 600 AI use cases across operations, with customer service applications spanning from automated inquiry handling to intelligent workflow automation.

Why It Works: Modern conversational AI leverages natural language processing (NLP) and machine learning to understand customer intent, context, and sentiment. These systems can handle routine queries instantly, check balances, review recent transactions, explain fees, and provide account information, freeing human representatives to focus on complex issues requiring more judgment. According to industry research, by 2030, voice assistants are expected to handle 80% of customer transactions in banking.

Despite progress, challenges remain. Deloitte’s 2025 survey found that 74% of banking customers still prefer human agents for routine queries, and 57% of chatbot users cited accuracy as the top area needing improvement.

2. Chatbots & Virtual Assistants for Smaller Institutions

AI-powered chatbots and virtual assistants are now foundational to customer service strategies across financial institutions of all sizes. While large institutions have built proprietary platforms, small to mid-sized institutions are increasingly adopting flexible, cost-effective AI solutions tailored to their scale and regulatory needs.

Examples of Tools in Action:

1. Posh Technologies: Designed specifically for credit unions and community banks, Posh offers conversational AI that automates member support across voice and digital channels. Institutions have used Posh to reduce call center volume and improve member satisfaction.
2. interface.ai: This platform provides AI-powered virtual assistants and intelligent call routing for credit unions and community banks.
3. Glia: Glia’s digital customer service platform includes AI-powered chat and voice assistants that integrate with core banking systems. It’s widely adopted by mid-sized institutions seeking to modernize their contact centers without overhauling existing infrastructure.
4. ChatRef: A newer entrant focused on credit unions, ChatRef offers customizable AI chatbots that handle member inquiries, loan support, and account services with 24/7 availability.

Why It Works: These platforms are designed with the unique needs of smaller institutions in mind, offering out-of-the-box integrations with core systems, compliance-ready features, and pricing models that scale with institution size. They enable smaller institutions to deliver the same level of digital service as national banks, without the need for in-house AI development.

How They’re Getting Better: AI chatbots are evolving beyond simple question-and-answer interactions. Modern systems incorporate:

• Context retention across multiple interactions, remembering previous conversations
• Proactive engagement, alerting customers to unusual activity or opportunities
• Multimodal capabilities, supporting voice, text, and visual interfaces
• Continuous learning, improving accuracy through machine learning feedback loops

3. Sentiment Analysis & Real-Time Monitoring

Financial institutions are deploying AI-powered sentiment analysis to understand customer emotions during interactions, identify dissatisfaction early, and improve service quality.

Applications:

• Call center monitoring: AI analyzes tone, word choice, and conversation patterns during customer calls to detect frustration, confusion, or satisfaction.
• Chat and email analysis: Systems evaluate written communications to identify customer sentiment and urgency.
• Proactive intervention: When AI detects negative sentiment, it can alert supervisors for immediate intervention or escalation.

Real-World Impact: Sentiment analysis in financial services helps institutions:

• Identify emerging issues and churn signals in real time
• Reduce churn and improve customer satisfaction scores
• Prioritize fixes faster by understanding root causes
• Track sentiment over time during earnings calls and customer interactions

4. Hyper-Personalization & Recommendation Engines

AI-driven personalization has evolved from basic segmentation to hyper-personalization using real-time data, behavioral analytics, and machine learning to deliver individualized experiences at scale.

Examples:

• Personalized product recommendations: AI analyzes spending patterns, life events, and financial goals to suggest relevant products like credit cards, loans, or investment options
• Customized financial advice: Virtual assistants provide tailored budgeting guidance, savings goal tracking, and spending insights based on individual customer behavior

Challenges and Risks: What Financial Institutions Must Navigate

While AI offers significant benefits in customer service, financial institutions face substantial challenges in implementation and ongoing management:

Regulatory Compliance & Consumer Protection

The Consumer Financial Protection Bureau (CFPB) has made clear that AI systems must comply with all existing consumer protection laws. In a June 2023 report on chatbots in consumer finance, the CFPB highlighted concerns about poorly deployed chatbots impeding customers from resolving problems.

Compliance Requirements:

• Transparency: Financial institutions must provide clear disclosures about AI usage in customer interactions
• Accuracy: Chatbots cannot provide misleading information or make false promises
• Accessibility: AI systems must remain accessible to all customers, including those with disabilities
• Human escalation: Customers must have clear paths to reach human agents when AI cannot resolve issues

Data Privacy & Security

As AI systems process vast amounts of personal and financial data, privacy and security risks intensify. Financial institutions must implement:

• Bank-grade encryption for all customer data processed by AI systems
• Consent management ensuring customers understand and approve data usage
• Data minimization practices, collecting only necessary information
• Breach prevention through advanced cybersecurity frameworks

Customer Trust & Acceptance

Despite technological advances, customer trust remains a significant barrier. Key concerns include:

• Accuracy and Reliability: Among Deloitte survey respondents who preferred chatbots, 57% said better accuracy was the improvement they’d most like to see. Providing 100% accurate responses remains challenging due to AI hallucinations, instances where AI systems generate plausible sounding but incorrect information.
• Preference for Human Interaction: Deloitte’s 2025 survey revealed that 74% of respondents favored human agents over chatbots for routine interaction. This preference highlights a generational divide, with millennials and Gen Z reporting more positive chatbot experiences than baby boomers and Gen X.
• Trust in AI for Financial Decisions: Only 27% of consumers trust AI for financial advice, according to separate research cited in the Deloitte report. Banks must work to build trust through transparency, education, and proven reliability.

Bias & Fairness

AI systems trained on historical data can perpetuate or amplify existing biases. In customer service applications, this manifests as:

• Unequal service quality across demographic groups
• Biased routing that directs certain customers to inferior automated systems while others reach premium human agents
• Discriminatory treatment in credit-related inquiries or product recommendation

Financial institutions must implement bias detection tools early in the design process and use fairness-aware algorithms to ensure equitable treatment. Wells Fargo, for example, improved its AI credit scoring models and reduced racial bias in loan approvals by 25% through deliberate bias mitigation.

Integration with Legacy Systems

Many financial institutions struggle to integrate AI tools with existing core banking platforms, customer relationship management (CRM) systems, and data warehouses. Seamless integration is critical for AI to access the data it needs and deliver value.

Integration Challenges:

• Data silos preventing AI from accessing complete customer information
• Outdated APIs that don’t support modern AI platforms
• Real-time processing requirements exceeding legacy system capabilities
• Compliance constraints limiting data access and movement.

Practical Guidance: Implementing AI in Customer Service

For financial institutions considering AI adoption in customer service, industry experts and early adopters recommend the following approach:

1. Start with High-Impact, Lower-Risk Use Cases

Begin with AI applications that deliver clear return on investment and have established implementation patterns:

Recommended Starting Points:

• FAQ chatbots handling common inquiries about hours, locations, and basic account services
• Password resets and account unlocking through secure automated flows
• Transaction alerts and notifications using AI to detect unusual activity
• Customer feedback analysis using sentiment analysis to identify service issues

These use cases allow institutions to build confidence, develop internal expertise, and establish governance processes before tackling more complex applications.

2. Prioritize Customer Education & Transparency

Educating customers about AI builds trust and improves adoption. Financial institutions should:

• Clearly identify AI interactions, so customers know when they’re chatting with a bot versus a human.
• Explain AI capabilities and limitations upfront to set appropriate expectations.
• Provide educational content through articles, videos, and infographics explaining how AI protects and serves them.
• Offer choice by allowing customers to opt for human assistance at any time.

Bank of America’s approach with Erica demonstrates this strategy, the bank provides extensive educational materials helping customers understand AI’s role in managing finances.

3. Build Human-AI Collaboration, Not Replacement

The most successful AI customer service implementations treat AI as augmentation rather than replacement:

Best Practices:

• Intelligent routing that directs simple queries to AI while reserving complex, emotional, or sensitive cases for human agents.
• Agent assist tools that provide human representatives with AI-generated suggestions, relevant information, and next-best-action recommendations.
• Seamless handoffs when AI reaches its limits, transferring context and conversation history to human agents.
• Continuous improvement using human agent insights to train and refine AI systems.

Deloitte research emphasizes that banks should design seamless collaboration between human agents and AI, improving the precision and contextual relevance of responses and actions.

4. Establish Robust Governance & Ethics Frameworks

Safe AI adoption requires strong governance structures that includes:

• AI ethics committees composed of data science, legal, compliance, and customer experience experts.
• Explainable AI (XAI) architectures that can clarify how decisions are made.
• Continuous monitoring for bias, accuracy, and compliance.
• Clear accountability with defined roles for AI oversight and incident response.

5. Choose the Right Technology Partners

Vendor selection is critical for successful AI implementation. Financial institutions should consider the below selection criteria:

• Security and compliance certifications (SOC 2, ISO 42001, etc.)
• Integration capabilities with existing systems and platforms
• Explainability and transparency in AI decision-making
• Customization options to match institutional needs and brand
• Scalability to grow from pilot to enterprise-wide deployment
• Support and training for implementation and ongoing maintenance

6. Implement Phased Rollouts with Continuous Learning

Rather than attempting enterprise-wide transformation immediately, successful institutions adopt phased approaches.

Implementation Phases:

• Pilot: Deploy AI to a limited customer segment or specific use case
• Measure: Track key metrics including accuracy, customer satisfaction, cost savings, and issue resolution rates
• Refine: Use insights to improve AI performance, expand training data, and optimize workflows
• Scale: Gradually expand to additional use cases and customer segments
• Optimize: Continuously monitor and improve based on performance data and customer feedback

Google Cloud’s research on AI ROI in financial services found that 53% of financial services executives report their organizations are leveraging AI agents in production, with 49% allocating over 50% of their future AI budget to AI agents.

Looking Ahead: The Future of AI in Customer Service

As we move through 2026 and beyond, several emerging trends will shape the future of AI-powered customer service in financial institutions.

The banking industry is entering what Lloyds Banking Group calls “the agentic moment”, a shift from reactive AI chatbots to proactive AI agents that can autonomously observe, plan, and act toward goals. According to BCG research, AI agents already account for 17% of total AI value in 2025 across all industries and are expected to reach 29% by 2028.

However, BCG also notes that only 5% of companies are currently generating value at scale from AI, with 60% not yet realizing significant benefits. The gap will widen as AI-first institutions gain competitive advantages that become difficult for others to overcome.

For financial institutions just beginning their AI customer service journey, the message is clear: start now, start small, but start strategically. The future of customer service is already here; the question is how quickly your institution will adapt to deliver it.

If you have any questions or are interested in learning more, we are here to help. Please do not hesitate to reach out to discuss your specific situation.

Sources:

1. Deloitte (2025). “AI banking chatbots: from frustration to delight.”
2. OpenTools.ai. “JPMorgan Chase sets new benchmark with AI-powered megabank transformation.”
3. Consumer Finance. “Chatbots in Consumer Finance”
4. Potential.com. “How voice AI is reshaping customer service in banking and finance.”
5. NetApp PDF: “CSS-7217 FSI Use Case.”
6. SCIRP (Open Journal of Business and Management) PDF
7. Google Cloud. “ROI 2025: Finance” (PDF).
8. BCG. “2025 Retail Banking Report” (PDF).

This material has been prepared for general, informational purposes only and is not intended to provide, and should not be relied on for, tax, legal or accounting advice. Should you require any such advice, please contact us directly. The information contained herein does not create, and your review or use of the information does not constitute, an accountant-client relationship.