Banks and artificial intelligence: a technological revolution of opportunities and challenges

The AI revolution is already underway in the banking sector.

Artificial intelligence technologies are rewriting operating models, making risk management more efficient, customer service smarter, and internal processes faster.

Bank employees’ roles are also evolving: AI supports people by offering predictive and decision-making tools. But alongside opportunities, challenges are growing — especially in areas such as security, transparency, and governance.

A rapidly expanding phenomenon

According to a recent report by ABI Lab, 69% of Italian banks are already running AI pilot projects.

The most common applications include credit scoring, anomaly detection, and customer service via intelligent chatbots.

Generative AI is also gaining attention, with models capable of creating content, analysing complex data and supporting strategic decisions.

For banks, AI represents a concrete opportunity to innovate service models — improving efficiency, reducing costs, and delivering increasingly personalised experiences.

Financial impact: beyond cost reduction

Estimates suggest that AI could reduce the cost of risk and compliance controls by up to 60% by 2028, with a direct contribution to anti-money laundering, fraud detection, dynamic risk assessment and insurance fraud prevention.

In addition, the cost of using large language models is dropping by more than 50% per year, making AI investments increasingly cost-effective.

But according to many industry analyses, the most relevant benefit is direct revenue growth.

AI enables staff to offload repetitive, low-value tasks — freeing up time to focus on customer relationships, personalised financial advice and more effective sales activities.

It’s not just about efficiency, but tangible economic impact: GenAI adoption could account for up to 30% of revenue in financial services firms.

Governance and compliance: the transparency challenge

The adoption of AI in banking faces several infrastructure hurdles.

Many banks still run on legacy core banking systems — with an average age of around 30 years — and complex architectures that hinder integration of new technologies.

Beyond technology, regulatory and organisational factors are central.

Implementing AI requires formalised processes, proper controls and advanced governance models.

Transparency and security are key: AI systems must be explainable, monitorable and built on quality datasets to avoid bias and opaque decisions.

The upcoming EU AI Act reinforces these principles, imposing strict requirements — especially for high-risk use cases such as credit approval or behavioural monitoring.

Smarter services, but trust is essential

AI makes banking services faster, more accurate and more personalised.

It can anticipate needs, deliver automated advice and stop fraud in real time.

But none of this matters unless clients trust it.

A recent study found that 62% of customers would be willing to use a smart agent as their personal financial assistant.

However, they must know when they are interacting with an automated system, how their data is being used and what the technology’s limitations are.

No matter how advanced, intelligent systems are not infallible.

In banking — where every decision has impact — human oversight remains essential to ensure a responsible and visionary approach to the future.