AI Agents: The Next Big Disruption in Banking and Finance

AI agents are poised to reshape banking, finance, and investment management in ways we’re only beginning to understand. But first, let’s clarify what we mean by AI agents, and how they differ from traditional AI chatbots.

From Chats to Autonomous AI Agents

We’re all familiar with chatbots, the rule-based or LLM-powered assistants that engage in simple back-and-forth conversations. While useful, these systems are limited; they can answer questions but don’t truly act on behalf of users.

AI agents take things a step further. They have access to tools, external data sources, and the ability to execute multi-step processes, making them significantly more autonomous and capable.

For banking and finance, this shift is particularly exciting because so many tasks involve structured, multi-step processes that require analysis, decision-making, and execution. AI agents are well-suited to these challenges, and major financial institutions are already deploying them in ways that signal a coming revolution.

How Banks and FinTechs Are Using AI Agents

Capital One: AI Concierge for Car Buying

Capital One recently launched an AI-powered concierge designed to help customers navigate the car-buying process. Unlike a simple chatbot, this AI agent actively works with users to:

• Research vehicles

• Compare financing options

• Schedule test drives

• Book dealer appointments. 

This kind of hands-on assistance is a perfect example of how AI agents are not just reactive but proactive in helping customers make complex decisions.

Prem Natarajan, head of enterprise AI at Capital One, emphasized that the system was designed to mimic human reasoning, coordinating multiple AI agents to tackle different parts of the car-buying journey while ensuring accuracy and compliance. The company sees this as the foundation for broader AI-driven customer interactions in banking.

Goldman Sachs: AI Assistants for Investment Banking & Coding

Goldman Sachs has built an AI agent ecosystem called GS AI, which acts as an internal virtual assistant for thousands of employees. Currently, it supports investment bankers, analysts, and developers, streamlining workflows by retrieving data, generating reports, and even assisting with software development.

One of the most interesting use cases? AI agents are being developed to help bankers prepare M&A deal books, a process that traditionally takes hours of manual research. Goldman is also rolling out AI coding assistants for its 12,000+ developers, increasing productivity by 10-20% in time savings.

Goldman’s CIO Marco Argenti has likened AI’s impact to discovering a new energy source. He notes that the challenge isn’t just having the AI but also building the infrastructure and governance to harness it safely:

For now, Goldman requires AI agents to operate under strict human oversight and is proceeding cautiously.

Sardine: AI Agents for Fraud & Compliance

On the startup side, companies like Sardine are betting big on AI agents for risk and compliance. The company just secured $70 million in Series C funding to expand its AI-driven fraud prevention platform, which automates complex compliance tasks like:

• KYC (Know Your Customer) onboarding: Verifying identities and resolving discrepancies.

• Sanctions screening: Automatically checking global watchlists.

• Merchant risk assessment: Evaluating businesses for fraud risk before onboarding.

• Dispute resolution: Automating chargebacks and fraud claim investigations.

Sardine’s AI agents allow banks and fintechs to reduce false positives, speed up fraud detection, and improve compliance efficiency. As CEO Soups Ranjan notes, the volume of fraud alerts has grown 800% in recent years, and financial crime teams simply can’t keep up without AI assistance.

AI Agents in Private Credit & Investment Research

Another prime use case for AI agents? 

Private credit research and investment analysis. In this field, analysts spend hours digging into company financials, public filings, and market reports to compile deal memos. AI agents can automate much of this research, summarizing key financials, identifying risk factors, and even drafting investment overviews.

Fintech startup Auquan, for instance, has built an AI agent that reads earnings reports, news articles, and financial statements to generate credit memos for private lenders. Analysts still review and refine the AI’s output, but the process now takes minutes instead of days.

The Risks: AI Hallucinations, Compliance, and Regulation

Despite the promise, AI agents come with real risks, just like any other AI-driven system.

• Hallucinations & Errors: Even the most advanced AI models can generate inaccurate or misleading information.

• Regulatory Uncertainty: There’s still no clear legal framework for AI agent decision-making in banking, which means companies must be extra cautious.

• Risk Management & Oversight: Most firms keep humans in the loop to validate AI outputs—particularly in high-stakes areas like lending approvals, trading, and fraud investigations.

Goldman Sachs, for example, has explicitly barred AI agents from making autonomous financial decisions. Instead, it requires them to serve as assistants that provide recommendations, not execute trades or approve loans. This cautious approach will likely define early AI agent deployments in banking.

What’s Next? AI Agents Are Coming, But Carefully

The future of AI agents in finance will depend on three key factors:

1. Advances in AI models: As LLMs improve, AI agents will become more reliable and capable.

2. Stronger regulatory frameworks: Governments and financial regulators must set clear guidelines.

3. Enterprise AI adoption: Banks will continue to refine AI agents for safe and effective deployment.

In the near term, expect banks and fintech companies to prioritize AI agents for:

👉 Lower-risk, high-efficiency tasks

👉 Automating research

👉 Compliance checks

👉 Customer service

High-stakes decisions (like lending approvals and trading) will remain under human oversight for now.

💡But make no mistake: AI agents are changing the game. The financial industry is entering an era where AI is not just assisting but acting carefully, cautiously, and with an eye on the future.

About the Author

Manny Bernabe is an AI evangelist, LinkedIn Top Voice, and educator with a background in startups and finance. He has worked with AI startups and ran his own AI consultancy, helping businesses adopt emerging technologies. Previously, he held roles in the investment management industry, earning his CFA designation and a B.S. in Finance from DePaul University. Learn more about AI Agents in my video: AI Agents Explained in 7 Minutes.

🚀 Boost your business with us—advertise where 10M+ AI leaders engage

🌟 Sign up for the first AI Hub in the world.

📲 Our Socials