Article | August 14, 2026

Time for financial services AI to really earn its keep

By Ihyeeddine Elfeki, FSI Solutions Global Lead, DXC

For years, financial services institutions have treated artificial intelligence as an innovation portfolio: promising pilots, regulated experiments and board updates. That phase is ending.

The issue is now about practicalities: can AI improve customer outcomes, reduce operating drag and strengthen risk control inside institutions that are regulated, scrutinized and constrained by decades of accumulated systems?

In other words, can banks, insurers and market infrastructure firms turn AI into an operating advantage?

The appetite is clear enough. A 2026 Cambridge Judge Business School report found that 81% of surveyed financial services firms are adopting AI at some level (40% advanced). Yet only 14% see AI as transformational to organizational strategy and competitive advantage.

Many firms are active. Few have changed how they compete. 

Deloitte’s 2026 banking outlook makes the same point from another angle: many AI efforts remain stuck in isolated proofs of concept, slowed by dispersed data, compliance demands, legacy systems and an internal resistance to change. 

The constraint is rarely the model itself; it’s the operating environment around the model. A chatbot can’t fix a mortgage journey if customer data is scattered across channels. A fraud tool can’t build trust if decisions aren’t explainable.

AI value emerges when the business changes the process, accountability and customer experience around it.

Financial services companies have a different AI challenge from retail or media. Mistakes don’t merely irritate customers; they can also affect access to credit, claims outcomes, retirement decisions or regulatory standing.

Governance cannot be treated as a brake on innovation. It’s becoming the license to scale.

McKinsey warned that GenAI can expose financial institutions to legal and reputational risks while increasing vulnerability to cyberattacks and fraud. Similarly, the Financial Stability Board has highlighted concerns, including third-party dependencies, market correlations, cyber risks and model governance challenges.

Board-level questions are therefore sharper:

• “Which AI use cases touch regulated decisions?”

• “Who owns the outcome when an AI-assisted recommendation is wrong?”

• “Which vendors are becoming operationally critical?”

The strongest AI use cases improve an outcome executives already care about.

In banking, that may mean reducing digital drop-off or making services accessible to more customers. Banco Sabadell has been working with DXC to improve digital accessibility for 12 million customers in Spain, with a new framework using manual testing, automation and AI-based analysis to identify accessibility issues across web platforms. The initiative intends to reduce user drop-off, improve customer satisfaction and support greater independence in day-to-day digital engagements. 

Last year DXC also announced a 10-year AI-powered banking operations transformation with Spain's Unicaja bank, which wants to position this work as a global benchmark.

In customer-owned banking, the outcome may be increased personalized services without adding cost. One Australian mutual bank had employees navigating 15 screens and up to 26 clicks to piece together customer information. A 360-degree customer view gave staff consolidated access to customer history, accounts, correspondence and portfolio information. 

In insurance, the issue may be compliance and service quality together. A large Australian life insurer moved from a legacy contact center environment to a cloud-based engagement platform across Australia and New Zealand. The program delivered 24/7 access for agents and supervisors, was completed across four releases over 7 months, and enabled 100% call quality assurance.

These examples show AI being useful in context. The technology is embedded within accessibility, customer service, compliance and productivity.

Data modernization is a must

Every board wants AI-enabled personalization. A growing percentage want to discuss the data work that makes personalization safe, timely and profitable.

Yet, financial institutions can’t personalize based on unreliable data. They can’t automate controls around inconsistent records. They can’t explain model behavior when the information lineage is unclear.

Deloitte’s regulatory outlook for 2026 notes that as AI scales into more critical processes, supervisors are focusing more heavily on data governance and business continuity. High-quality data underpins transparency, model validation, explainability, fairness and oversight. 

This is why AI strategy and modernization strategy are now inseparable. AI should be a forcing function to simplify workflows, retire duplication and form clear paths between data, decisions and customers.

The next phase of AI in financial services won’t involve experimentation alone. Leaders should select a small number of high-value journeys, define the customer or risk outcome clearly and fund the operating change needed to make AI dependable at scale.

And their choice of domain-savvy technology partner is critical. DXC has 45+ years of financial services experience, supports 350+ active banking clients across more than 70 countries and maintains productive relationships with 17 of the world’s top 20 banks.

The key to success in the industry is to stop asking where AI can be tested and start deciding which outcomes need to be rebuilt. AI’s promise won’t be proven in the financial services lab. It’ll be realized in faster service, stronger controls, more inclusive access and operating models that keep on improving after their initial deployment.

Learn more

• AI Workbench
• Financial Services IT Solutions