From Black Box to Agentic AI: Reflecting on AI’s Adoption and Maturity

I have spent years working in financial services and have witnessed AI’s incredible transformation journey firsthand. Today, AI is not a future concept—it’s here and has brought immense change to financial services.

How is AI Transforming Financial Services?

Looking back at AI’s early days, financial services used AI for marketing and lead generation tasks. However, adoption was notably slower for financial crime prevention, not because of a lack of recognition of AI’s potential value. Instead, it typically stemmed from a genuine and valid need for caution. Financial institutions are inherently risk-averse, so there was a natural reluctance to fully rely on AI systems that, at the time, lacked a demonstrable track record of measured success.

This initial caution fueled a growing demand for greater control and transparency over AI investments. The industry began to push back against black box models that hid AI’s decisioning and demanded solutions that provided clear explainability, robust auditability, and the ability to own and fine-tune their AI systems to meet specific and changing needs. 

Putting Trust at the Heart of Innovation

This shift signified the industry’s move towards a more responsible and sustainable approach, where trust and understanding of AI are paramount. Rejecting black box systems signified an industry transition toward greater ownership and control over the AI solutions banks used. No longer simply purchasing a solution from a vendor. Your bank can now actively collaborate with technology providers to ensure their AI systems are tailored to your unique business requirements. 

Exerting greater control altered the nature of bank and vendor relationships. Banks are now partners in their AI initiatives, able to negotiate power and better understand a system’s performance and value. Ultimately, it empowers your financial institution to optimize the customer journey for a superior banking experience.

AI’s Evolution: From Reactive to Real Time

Another significant milestone in AI’s evolution and maturity is its incorporation into operations teams and ecosystems. AI was initially deployed as a reactive tool, most notably for fraud detection. These early systems excelled at identifying patterns and anomalies that pointed to fraud. However, there was a high risk the fraudster would ultimately disappear with the money. 

Fueled by advancements in machine learning, AI is now real time by design and capable of proactively identifying and preventing fraud where immediate decisions are paramount. By analyzing massive volumes of data at incredible speed, AI can swiftly detect suspicious activities and intervene to prevent financial losses.

Welcome to the Agentic AI Age 

Now we’re seeing adoption of agentic AI as a viable technology toolset rapidly accelerating. Financial institutions are on the cusp of deploying AI to proactively engage with their customers across a multitude of touchpoints, from AI-powered chatbots to automated claims management.

Advancements in agentic AI promise to significantly enhance the customer experience through 24/7 availability and personalized service. AI is transforming claims management by automating assessment, paperwork processing, and even payout determinations, dramatically accelerating processing times and reducing administrative overhead. 

Agentic AI should not displace human engagement. People must still be included in the loop to ensure the model’s outcomes match the customer’s intent. However, it can free human agents to focus on complex customer interactions and investigations. 

The agentic AI evolution demands a change in how banks acquire and develop talent. Update your hiring strategies to attract professionals with a strong understanding of AI governance, risk management, and compliance. In essence, start building teams that are AI-native. Your hires should be comfortable working alongside AI, understanding its capabilities and limitations, and capable of overseeing its responsible and ethical applications. 

It’s also time to cultivate the next generation of AI talent. Consider updating any apprenticeship programs you operate to recruit and train younger entrants into AI. By equipping people in their teens or early 20s with the AI skills they need for the future, you’ll improve your chances of recruiting candidates with the essential AI skills.

I firmly believe that 2025 will be remembered as the year agentic AI—the next chapter in AI maturity—becomes the norm. When this happens, it will fundamentally transform how banks and financial institutions operate and how customers experience financial services.

This isn’t a gradual evolution. It’s a dynamic revolution, and we’re already fully immersed in it.

• Blog: AI-Driven Fraud: How Criminals Are Using AI to Scam You
• Blog: How AI is Helping Fight Back Against Financial Crime
• Report: 2025 AI Trends in Fraud and Financial Crime Prevention
• Solution: RiskOps. Unify All Risk Signals to Stop FinCrime.