Financial Services Leaders Seek to Bridge the Gap Between AI Vision and Real-World Execution

Artificial intelligence (AI) promises to impact the banking/financial services/insurance (BFSI) industry across the board, transforming everything from risk assessment to customer engagement. But the industry's journey to AI is being slowed by a combination of inefficient processes, legacy infrastructure, regulatory complexity, and organizational inertia. While executives overwhelmingly recognize that digitization and AI adoption are existential imperatives, they are struggling to realize their potential.

A new survey of more than 500 senior BFSI executives conducted by HFS Research in partnership with Iron Mountain underscores the challenges their institutions face. Nearly 78% of executives said failure to digitize could result in permanent competitive irrelevance, yet only one in five firms is taking decisive steps to replace the traditional back office. Bold visions for AI-enabled transformation exist on paper, but most organizations remain mired in slow, incremental change.

In an industry characterized by repetitive processes, executives are optimistic that AI can shoulder many routine workloads. Over 80% predict autonomous AI agents will eventually handle at least three-quarters of everyday tasks, and 77% believe the traditional back office will vanish within three years. However, just 13% of their firms have deployed agents at scale, and only 21% say they are aggressively pursuing a goal of back-office transformation.

Outdated legacy systems and infrastructure are the number one impediment to progress, cited by 46% of executives. The industry’s continued reliance on paper also adds to the drag. Although 72% of executives rank digitization as a top priority, only one-third are confident they can securely digitize records at scale.

Realizing potential

There are good reasons BFSI executives are so riveted by AI’s potential. In addition to cutting the cost of labor-intensive processes, AI can be an engine for innovation, customer engagement and process reinvention. AI agents, which operate autonomously and in concert with each other, can cut processing times from days to minutes. Generative AI can surface insights from decades of archived records, enabling more accurate risk modeling and personalized financial products.

In a heavily regulated industry, AI can also be used to embed audit trails and controls into digital workflows, enabling financial services organizations to move from costly, reactive compliance to proactive, real-time monitoring. The survey found that regulatory compliance improvements are the number one objective driving back-office transformation, tied with efficiency and productivity gains.

Customer expectations are making the AI quest more urgent. The new cadre of Generation Z digital natives is less loyal than prior generations and expects high levels of automation to fuel self-service, investment advice and payments. Delivering these capabilities requires reimagining the back office as a growth driver rather than a cost center.

The survey identified several “hard truths” confronting BFSI executives’ AI ambitions.

Legacy infrastructure and technical debt are unfortunate byproducts of financial firms’ early forays into computing and are seen as the number one roadblock to achieving fully digital outcomes. Maintaining aging infrastructure ties up budgets that could be spent on innovation, and fragmentation hinders the collection of the high-quality data needed for AI training. Organizations need to adopt more flexible cloud platforms, modernize code and shed their reliance on brittle, siloed applications.

Paper-laden workflows are inefficient and risky. They block the transition to parallel workflows that digitization enables. They undermine AI initiatives that depend on structured, machine-readable data. Optical character recognition and generative AI now enable troves of printed documents to be digitized and understood. Yet while 72% of executives surveyed want to eliminate paper, only 34% feel confident they can do so while staying compliant.

Governance frameworks for AI are immature. Executives are understandably wary about the risk of bias, errors, and hallucinations; more than one-third said security and trust are barriers to achieving digital ambitions. AI efforts require dedicated governance playbooks that specify how decisions are explained, lay out ethical and accountability guidelines, and clearly define the role of human oversight.

Talent shortages loom. An impending wave of retirements will cost the insurance industry alone about 400,000 workers by 2026. Those who remain need to be re-skilled as supervisors of and collaborators with AI systems. Nearly half of executives believe AI will require significant employee upskilling. The good news is that most roles will be repositioned as more strategic.

Playbook for success

Overcoming these obstacles requires changes to technology, governance, and culture.

• That starts with moving beyond traditional metrics like cost savings to measure success in terms of improved trust, compliance, and customer experience.
• Organizations need to get their data house in order. AI systems trained on poor-quality data are unreliable and even dangerous. Institutions need to modernize data governance, adopt open standards, and unify information silos.
• Compliance processes have too long been layered in after the fact. AI enables real-time monitoring and embedded audit trails that reduce cost and risk.
• Training programs should prepare employees to treat AI agents as digital coworkers, using them to redefine processes and create new customer value. A collaborative approach to AI adoption transforms fear of replacement into a culture of partnership.
• Shed the “not invented here” mindset and get involved with collaborative consortia, regulators, fintech partners and forward-looking vendors to set standards, share costs, and accelerate safe adoption.

The payoff

Citigroup has estimated that AI could add as much as $2 trillion to global banking profits by 2028. In a market where fintechs and neobanks are eroding customer loyalty, institutions that can offer seamless, AI-powered experiences are better positioned to acquire and retain the new breed of digital native customers.

“We need to stop optimizing 20th-century processes and start building 22nd-century capabilities,” said one survey respondent. “The dimensional shift requires abandoning improvement for total reconstruction.”