Agentic AI Pushes Banks to Fix Security, Data and Decision Rights

And one of the biggest issues there is in identifying non-human identities, or as they’re better known, bots. Security Boulevard argues that agentic AI is starting to reshape financial services, but only if institutions strengthen the security systems that sit underneath it. The article centers on “non-human identities,” or the digital credentials that let applications, servers and automated systems interact with one another. It says these machine identities are becoming more important as banks and financial firms move more of their operations to the cloud and rely more heavily on automated tools. In that setting, better management of passwords, tokens and access rights can help reduce risk, improve compliance, give security teams more visibility and lower operating costs.

The article also says agentic AI can help financial institutions make faster decisions, improve fraud detection, streamline operations and support new products and services. But Security Boulevard stresses that these gains depend on tighter control over which systems can access sensitive data and how they behave once they do. It presents security culture, automation and closer coordination across teams as essential parts of that work. The main takeaway is that agentic AI may open new opportunities in financial services, but firms will need stronger machine identity management and a more unified cybersecurity strategy to use it safely.

Context Engineering

That safety is supported by an inside baseball term called context engineering. Context engineering is the work of giving an AI system the right business information, rules and goals so it can make useful decisions in the right setting. In banking, that means grounding AI tools in a bank’s own data, policies and customer context so their responses are more relevant, more reliable and less likely to go off track.

Financial IT says this step is essential if banks want autonomous AI to move from simple experiments to real operational use. The article says banks are already using these tools to prepare client materials, model financial scenarios, speed outreach and improve fraud detection. It also argues that context-rich AI can help reduce manual work, improve employee productivity and support faster, more tailored customer service.

Financial IT also says the next phase of autonomous AI in banking will depend on governance, compliance and the strength of the underlying data foundation. The article points to examples of banks and payments companies using AI to cut onboarding time, reduce false positives in anti-money laundering checks and strengthen fraud controls by adding more context to transaction analysis. It also says banks could use autonomous AI to deliver highly personalized financial guidance and new embedded banking services.

But the piece stresses that these gains will only be sustainable if firms build systems that are auditable, secure and aligned with rules such as the EU AI Act. The main message is that autonomous AI can improve speed, service and decision-making in banking, but only when it is grounded in trusted data and deployed with strong oversight.

Message to CMOs

Banking CMOs are also taking a look at agentic AI, and it was the topic of a Wall Street Journal article last week. The article argues that agentic AI could become a major force in how banks market to customers, personalize offers and shape growth. It says some banks are already using these systems to monitor campaigns, adjust messaging and create tailored outreach based on signals such as payroll deposits or travel spending.

The article also points to uses in investment banking and wealth management, where AI can help produce research commentary and personalized client reports more quickly. Its broader point is that agentic AI is moving beyond back-office efficiency and starting to influence how banks make commercial decisions.

The article’s main warning is that many bank CMOs may not yet control the functions that matter most in an AI-driven environment. Deloitte’s analysis of the 28 largest U.S. banks found uneven CMO influence over customer data, analytics, marketing technology and customer experience, while privacy and consent management often sit elsewhere. The authors say that if marketing leaders do not help shape the data, rules and incentives behind AI systems, they risk losing strategic influence. The takeaway is that CMOs should work more closely with technology, risk, finance and senior leadership so they can play a larger role in guiding AI strategy and turning agentic systems into a growth tool for the bank.

For all PYMNTS AI coverage, subscribe to the daily AI Newsletter.