Banks Adopt Digital Shift AI Next Frontier
Publicis Sapient, a global technology consultancy founded in 1990, has built its reputation by turning technical advancements into practical business solutions. Operating across 72 countries with over 20,000 employees, the company’s most transformative work has come from helping businesses adapt to major changes in customer engagement strategies.
Dave Murphy, the firm’s Head of Financial Services for EMEA and APAC, joined in 1997—a period when the World Wide Web was fundamentally altering industries. The early internet didn’t merely serve as a tool; it forced companies to rethink their entire operations. “It wasn’t just the introduction of the World Wide Web itself but the moment companies and corporations started to adopt it that fundamentally changed the way they engaged their customers and thought about their operations,” Murphy says. “Because you could no longer hide the rest of your company behind a human, especially in the case of banks or similar industries. People wanted direct interaction, all the way to the back office, in order to serve themselves.”
This transformation took time. Digital adoption among traditional banks reached only 35–40% by 2014–2015. However, the pandemic accelerated the trend dramatically. Today, 80–90% of all customer interactions with banks occur digitally—a shift that even the oldest institutions could no longer ignore.
While banks now function as digital-first organizations, the transition has not been without challenges. Murphy highlights new vulnerabilities introduced by digital-native operations. “Previously you were protected, because even if there were huge technological shifts happening in the industry, you could still serve your customers through human touch, the branch or the telephone,” he explains. “But now that you’re a digital business, as those technologies change, you have to change with them.”
AI represents the next major disruption. Murphy distinguishes between two stages: machine learning, which has long supported personalization and data-driven revenue, and the emerging field of agentic AI. Early chatbots were limited in capability, but today’s agentic systems are being tested for live, end-to-end transactions, though only after implementing strict safeguards.
Sapient began integrating machine learning platforms into UK high street banks around 2016–2017, allowing data scientists to refine models incrementally. Agentic AI, however, demands a different strategy. “Banks need to think carefully about how they introduce that to customers,” Murphy says.
For fintechs, AI adoption is not just a matter of implementation but of approach. Some use it as an assistant, achieving efficiency gains of up to 15% through features like autocomplete. Others redesign entire workflows, deploying autonomous agents to handle tasks independently. The key difference lies in contextual application.
The potential rewards are significant, but the hype must be tempered with realism. “Expertise remains irreplaceable,” Murphy stresses.
Agentic AI and Back-Office Efficiency
The most immediate impact of agentic AI is in areas previously reliant on human intervention. Mortgage approvals, once a weeks-long process involving manual document reviews, now see AI agents verifying credit scores, income statements, and property valuations in real time. A pilot at a major UK bank cut underwriting times by 60% while maintaining compliance, though human oversight remains available for complex cases.
The Role of Humans in an AI-First Era
Despite concerns about job displacement, Murphy argues AI’s greatest value lies in enhancement, not replacement. In wealth management, AI now handles routine tasks like portfolio rebalancing and tax-loss harvesting, allowing advisors to focus on high-net-worth clients. Client retention improved only when advisors used the extra time for personalized financial planning.
The transition also requires new skills. Banks are retraining employees in AI literacy, but demand for specialized roles, such as prompt engineers or AI ethics auditors, outpaces supply. Fintechs are hiring aggressively, often recruiting from tech companies. The talent shortage is most acute in emerging markets, where digital infrastructure exists but skilled workers do not.
Fintech Choice: Helper Tools or Autonomous Agents
Generative AI can act as a simple helper, completing typed input much like an autocomplete feature. It can also be built into agents that execute tasks on their own. Selecting the helper model keeps the technology in a supportive role, while the autonomous model requires redesigning the entire workflow. The distinction influences how quickly firms see value from the technology.
