• Agentic AI shifting focus of prompt economy toward financial services
• Machine identities emerging as critical cybersecurity risk
• Stronger access controls improve visibility and reduce exposure
• AI enabling faster decisions and enhanced fraud detection
• Context engineering ensures AI operates with relevant data and rules
• Banks using AI for onboarding, AML, and customer engagement
• Governance and data quality key to scaling autonomous systems
• EU AI Act and regulation shaping deployment requirements
• Marketing and revenue functions increasingly influenced by AI
• Cross-functional alignment critical to managing AI-driven risk 

Developments in what has been described as the “prompt economy” are increasingly shifting from consumer and infrastructure concerns toward financial services, as banks begin to explore how agentic AI can be embedded across both offensive and defensive aspects of their business models.

At the center of this shift is a growing recognition that automation is no longer limited to discrete tasks. Instead, autonomous AI systems are starting to influence how banks execute operations, manage risk, and interact with customers.

This transition is driving new use cases, but also exposing new vulnerabilities.

One of the most pressing challenges lies in the rise of so-called non-human identities.

These machine identities, which include the credentials used by applications, servers, and automated systems, are becoming critical as banks expand their reliance on cloud infrastructure and interconnected digital tools.

As these systems proliferate, the ability to manage passwords, tokens, and access rights is emerging as a core risk management priority.

Poor oversight of machine identities can create blind spots in cybersecurity frameworks, while stronger controls can improve visibility, reduce operational risk, and support compliance requirements.

The growing adoption of agentic AI is amplifying both the opportunity and the threat. On one hand, banks are using these systems to accelerate decision-making, enhance fraud detection, and streamline internal processes.

On the other, the complexity of autonomous systems increases the risk of unauthorized access, unintended behavior, and systemic vulnerabilities.

Industry analysis suggests that the benefits of agentic AI will only be realized if institutions tighten control over how systems access sensitive data and how they behave once deployed.

This requires not only technical safeguards but also a broader shift toward stronger security culture, greater automation in monitoring, and closer coordination across technology, risk, and compliance functions.

Alongside these developments, a concept known as context engineering is gaining traction within banking technology circles.

The term refers to the process of equipping AI systems with the appropriate business data, rules, and objectives so that they can operate effectively within specific organizational environments.

In practice, this means embedding AI tools within a bank’s internal data ecosystem and governance frameworks, ensuring that outputs are relevant, reliable, and aligned with policy.

Without this contextual grounding, autonomous systems risk generating inaccurate or inconsistent outcomes.

Financial institutions are already applying these approaches across a range of use cases. AI systems are being used to prepare client materials, model financial scenarios, support outreach, and enhance fraud detection. The addition of richer contextual data is also helping to reduce false positives in anti-money laundering processes and improve onboarding efficiency.

However, the transition from experimentation to full-scale deployment depends heavily on governance and data quality.

Autonomous AI systems must be auditable, secure, and aligned with regulatory expectations, including emerging frameworks such as the EU AI Act. Without these safeguards, the risks associated with automation could outweigh the benefits.

The implications extend beyond operations into commercial strategy. Marketing leaders are increasingly exploring how agentic AI can be used to personalize customer engagement, optimize campaigns, and drive growth.

Some institutions are already deploying AI to monitor marketing performance, adjust messaging in real time, and tailor offers based on behavioral signals such as spending patterns.

In investment banking and wealth management, similar technologies are being used to generate research insights and customized client reports more efficiently.

These developments highlight how AI is moving beyond back-office functions and becoming a driver of revenue generation.

Yet this shift also raises questions about organizational control and accountability. Analysis of large U.S. banks suggests that marketing leaders do not always have oversight of the data, analytics, and technology systems that underpin AI-driven decision-making.

Privacy, consent, and customer data management often sit in separate functions, creating potential gaps in governance.

As a result, there is growing recognition that successful AI adoption requires closer collaboration across departments.

Marketing, technology, risk, and finance teams must align around shared data standards, governance frameworks, and strategic objectives.

The broader message is clear. Agentic AI is opening new opportunities across financial services, from operational efficiency to customer engagement. But those opportunities come with heightened risk.

Banks that fail to strengthen machine identity management, data governance, and cross-functional coordination may find that the same technologies designed to drive growth also introduce new forms of vulnerability.