AI banking: How retail banks can use AI agents and even agentic AI to navigate uncertainty

We live in a moment of profound uncertainty. The instability of the political, economic and regulatory climate under the Trump administration is creating unprecedented challenges for banks trying to maintain normal operations.

AI banking tools could help financial institutions better navigate choppy seas. But banking leaders should look beyond generative AI to experiment with AI agents and even agentic AI.

AI agents can largely automate a specific task. Agentic AI is an even more powerful tool designed to operate with more autonomy to solve problems with little human oversight. One study found that banks using AI agents and agentic AI achieved up to 30% cost savings and 20% revenue growth over baseline.

However, fewer than one in four banks have woven this advanced AI tech into their strategic playbooks. As market pressures continue to ramp up, explore these five questions to consider how to develop a stronger AI strategy for your banking organization.

1. How could banks benefit from using advanced AI internally?

AI agents and agentic AI could offer your bank the opportunity to not just to find efficiencies but also change how you do business on a deeper level. This tech could transform entire workflows, automating certain operations and even taking certain responsibilities largely off the plates of your human team.

Using advanced AI can help your organization remain nimble and respond faster to rapidly changing market conditions. Here are a few areas in which bank leaders should explore agentic AI to improve internal processes:

• Know your customer (KYC): AI agents could be set up to ingest documents, extract data, run watch-list checks and prepare escalation reports with minimal oversight.
• Loan origination and servicing: Try experimenting with agentic AI to assemble credit files, calculate risk metrics and draft approval memos.
• Compliance and risk: Ask an AI agent to continuously watch for regulatory changes, flag them and create summaries to help keep your team updated on the implications.
• Autonomous transaction surveillance agents: An AI agent could triage alerts, helping your team prioritize which transactions to focus on first, plus draft investigation summaries.
• Finance and treasury: AI agents could allow you to automate intraday liquidity management, forecast cash needs, execute low-risk funding moves, manage journal entry processing and account reconciliations, draft financial statements and tackle call report prep.
• Autonomous liquidity management (ALM): An agentic AI could not just run scenario analyses here but also propose hedge adjustments when needed.
• IT support: Set up an agentic AI to handle low-level tech support incident remediation, execute run books and loop in engineers for exceptions, as well as answer employee questions.

2. How might banks use AI agents or agentic AI for customer-facing tasks?

One benefit of advanced AI is that it can handle certain people-facing roles that involve directly engaging with your customers. Here are a few specific areas you could experiment with automating:

• Offer virtual financial planning assistants that serve your customers by acting as a personal CFO to monitor account balances, execute bill pay and more.
• Use AI to negotiate or renegotiate customer rates when customers apply for loans or lines of credit.
• Build agentic robo-advisors to rebalance customer investment portfolios based on market moves, tax-loss harvesting windows and changing financial goals.
• Create conversational banking agents that allow customers to ask for help completing key tasks like “open an account” or “find me the best rate.”
• Support customer security by contacting them directly with proactive risk alerts or when fraud is detected.

3. How could banks begin experimenting with AI agents or agentic AI?

So far, you’ve gotten over a dozen different ideas on how you might use AI agents or agentic AI to strengthen your bank. That’s a lot to implement all at once, especially when you’re still experimenting with new tech.

So, where should you start using advanced AI first? One area to try early on is autonomous liquidity management (ALM). Banks have begun using agentic AI tools to ingest cash flow data and then propose overnight repo trades to optimize funding costs.

In some cases, the agent even executes overnight trades itself. And while it’s early days, results show a 25% reduction in liquidity shortfall events.

Credit underwriting and loan processing are also promising. Banks are building end-to-end lending agents that have cut processing times by up to 80% and even deploying autonomous call center agents to handle customer inquiries simple enough that they don’t need escalation to a human being to solve.

Autonomous agents are also proving useful for transaction reconciliation and bank-office ops. Here, AI agents are taking over up to 90% of routine reconciliations at banks willing to experiment with the technology, matching payments against invoices and ledger entries to slash both time on task and error rates.

4. What could the future look like for agentic AI in financial technology?

As agentic AI in banking evolves further, keep an eye on autonomous personal finance. Think of a personal AI concierge for every customer.

This AI concierge would be a single app on each customer’s phone that has access to all financial accounts and uses that information to provide holistic goal setting, continuous monitoring and execution, proactive recommendations and seamless handoffs with human interface when needed.

If this sounds farfetched, consider that consumers already want just one financial provider that can serve all their needs. So, if your bank can provide that service to become your customers’ central financial hub, you win.

Again, it’s early yet, but data shows that banks that enable autonomous services for their customers show 20%-30% higher product retention than competitors. In other words, banks that deploy AI agents to handle tasks for their customers get stickier customers.

5. What risks could agentic AI create for banks?

AI is not a panacea. The technology is still an unknown quantity in many ways and can pose significant risks.

Here are some core challenge areas:

• Lack of quality data: Like any other AI system, the value of an agentic AI is determined by the quality of data it has to learn from. In other words: Garbage in equals garbage out. While hallucinations, or an AI making up answers out of thin air, are a risk in any AI system, an AI fed on low-quality data may be more likely to hallucinate. Immature data governance can also result in a lack of trust in data.
• Siloed systems: Most banks operate on dozens of siloed core systems, which means an agentic AI needs to be capable of interfacing with each individual system.
• Maintaining data privacy and cybersecurity: Agentic systems often need both read and write access to customer accounts, which means that a compromised agent could expose or corrupt highly sensitive data.
• Auditability: Regulators demand clear, step-by-step accounting of risk transparency or compliance decisions, which can be more complicated to produce when an AI is involved in decision-making.
• Governance and control: Every agentic AI needs human oversight, along with guardrails and kill switches in place to prevent an AI from autonomously executing unintended transactions. You don’t want to wake up one morning to discover your AI has placed a $1 billion bet on a tulip bubble.
• Use complexity: The more complex the task, the more likely it is that even advanced AI will fail. AI tech can handle linear activities, but can’t match human reasoning. Here, you need to focus on fine-tuning your AIs, testing multiple scenarios and maintaining human monitoring, especially when letting an agentic AI run anything.

How Wipfli can help

Ask Wipfli to help you discover how you can best use advanced AI to strengthen your bank. Our team understands the nuances of banking regulations and how to integrate AI into your internal and customer-facing operations to improve your business performance. Learn more here.

Read more

• How tariffs impact financial institution risk projections
• How AI is reshaping data privacy
• Generative AI is revolutionizing the banking industry