Why the Finance Industry Is Turning to Agentic AI
Why the finance industry is looking – Imagine a future where an AI assistant manages your daily purchases. This is agentic AI—software capable of executing intricate tasks independently, not just responding to simple questions. At the Money 20/20 Europe event in Amsterdam, Mastercard, ING, and Worldline unveiled “Europe’s first live end-to-end agentic payment” last week. A consumer instructed an AI to search for concert tickets in a specific location on a set date, within a fixed price range. The assistant identified suitable options, processed the payment after selection, and required human verification to finalize the transaction.
AI Adoption Expands Beyond Startups
The conference, marketed as the most comprehensive annual event in financial technology, highlighted agentic AI as a transformative trend. For years, fintech firms and traditional banks were competitors, but recent collaborations show a shift toward shared innovation. Scarlett Sieber, the event’s chief strategy and growth officer, noted that AI in finance has evolved from a marketing term to a widespread reality. “Real-world implementation is now happening across all sectors,” she explained.
According to a 2026 University of Cambridge study, agentic AI deployment in the financial industry is projected to rise from 24% to 81% by 2030. The report, which analyzed data from over 600 institutions and regulators, warned that the pace of AI advancement often exceeds existing oversight systems. “The rapid evolution of technology currently outpaces the frameworks and technical capabilities needed to monitor it,” the study emphasized.
Real-World Applications and Challenges
Israeli company eToro showcased its AI capabilities with an app that monitors Donald Trump’s social media activity. When the president shares content potentially affecting markets, the AI can execute trades in user accounts within seconds. CEO Yoni Assia revealed that AI use within the firm has grown tenfold in six months, with 95% of new code now generated by AI. Yet, he stressed that “AI is useless without humans steering it.”
“If you put AI on a bad process, you still have a bad outcome,” said Marguerite Bérard, CEO of ABN AMRO, the third-largest Dutch bank. “Human oversight remains critical.”
Meanwhile, Swedish fintech firm Klarna integrated an AI assistant with OpenAI to handle customer service, claiming it replaces the work of 700 full-time employees. The company’s workforce has shrunk from 6,000 to fewer than 3,000 in recent years, while revenue per employee increased. However, Klarna’s CEO, Sebastian Siemiatkowski, admitted that cost-cutting had led to “lower quality” in some areas. The firm has since begun rehiring human agents, “investing in the quality of the human support,” he noted.
Industry-Wide Impacts and Concerns
As agentic AI becomes more prevalent, concerns about its broader implications are emerging. Gartner predicted that over 40% of AI projects would be abandoned by 2027, citing rising costs, uncertain value, and insufficient risk management. A report by Accenture and Wharton business school urged leaders to define how human judgment and accountability should be integrated into AI systems.
ABN AMRO exemplifies this trend, having reduced its physical branches from 500 in 2010 to just 26 today. The bank plans to cut 5,200 positions by 2028, with 85% of staff already utilizing AI tools like “Ana” for customer interactions and “Lenny” to streamline credit processes. Despite these advancements, the balance between automation and human oversight remains a central debate.
