Across financial services, the shift from process automation to decision-grade autonomy is no longer theoretical. Leaders want systems that can “think, act and decide on behalf of the human…to actually elevate the role of the human,” says Shyamala Chandrasekar, Solutions Principal, Autonomous Financial Services at HCLTech.

Recent research, Future of Payments: AI everywhere. Trust nowhere?, backs up the urgency: more than half of payments organizations expect to operate as autonomous financial services entities within two years, yet only a minority have the infrastructure and governance to do so confidently.

What autonomous really means in financial services

Autonomy isn’t about machines replacing people. As Chandrasekar puts it, “when you look at autonomous financial services, what we are really talking about is intelligent systems that can think, act and decide on behalf of the human…think about it as augmenting human judgment.” She likens it to aviation’s long-standing autopilot: pilots didn’t disappear; their roles became more strategic.

Under the hood, she says, autonomy blends large-scale automation with AI “as the brain,” and agents that orchestrate end-to-end workflows. All of it “sits on the foundation of composable architecture,” with data as the fuel and a human-in-the-loop to maintain trust. This frees up staff “to spend more time building relationships” and deepening conversations with customers.

The roadmap: Architecture and journey re-engineering

Progress will be staged. Chandrasekar outlines two levers. First, architectural maturity: data, composability, APIs and microservices. Second, journey re-engineering: reimagining end-to-end experiences so that AI doesn’t just digitize steps but “eliminate things that you don’t anymore need,” with intelligent systems acting on the user’s behalf.

That emphasis on foundations echoes the research. Only 20% of companies have fully modernized, real-time data systems and 81% lack fully modernized, cloud-native real-time capabilities, which represent significant challenges for autonomy at scale.

Agent-pay in practice: From event-driven payments to orchestration

Payments is fertile ground because, as Aaditya Rathod, Head – FinTech, UK & Europe at HCLTech, notes, it “cuts across each and every industry.” He defines autonomous payments as event-driven transactions “initiated and executed by systems and devices…on the basis of predefined rules,” powered by data and “a full AI engine.”

Organizations are getting ready. While only 18% say they are fully prepared to deploy secure agent-pay solutions in customer-facing operations, a further 63% are mostly prepared and closing gaps.

Rathod stresses the connective tissue of “robust APIs” and secure connectivity inside and outside the enterprise, as well as an orchestration layer so “the right messaging can flow into the right vertical.”

Trust, risk and governance: Confronting leaders’ biggest concerns

Caution is rational. “Trust and security are the core,” says Rathod. Nearly all executives share concerns about applying AI to payments (91%) and specifically about agentic assistants (99%), with top risks including privacy/security (42%), integration with legacy systems (40%) and customer trust (40%).

Rathod points to three immediate moves: adopt a zero-trust stance where “every transaction should be seen as new”, harden data protection and ensure regulatory readiness from the outset. The research underscores why these matter: 60% of leaders say today’s AI fraud tools are more ineffective than effective and 47% lack formal AI policies or guidelines, which are gaps that heighten the chance of misuse or malfunction.

Early value pools: Fraud, routing, investigations and beyond

Rathod adds two tangible use cases:

• Predictive treasury and liquidity: Corporate treasurers “sit on a huge amount of funds”. With autonomous engines, “the system can find out when the liquidity is needed” and automatically place surplus into interest-bearing accounts, then pull it back when required
• Liquidity-aware bill payments: Instead of fixed-date debits, consumers can instruct an agent to pay “whenever I have enough liquidity…bearing in mind the last date,” maximizing cash on hand until the due date

These patterns align with market pressure: 87% of executives fear losing customers without real-time capabilities, driving adoption of instant and event-driven experiences.

Investing without lock-in: Timing, scale and ecosystem plays

Half the battle is strategic patience. While 54% view AI agents as a long-term investment, 40% believe many current tools won’t be relevant in five years.

Rathod’s guidance is to get the security stack right, including data and infrastructure, be regulation-ready and elevate the user experience so customers always have “the right information…at the right time” for financial wellbeing.

Chandrasekar’s final advice: “this is an ecosystem play.” It’s not a do-it-yourself sprint but a coordinated effort across partners, front-to-back, with human-in-the-loop controls preserved as capabilities scale.

Embracing the autonomous future

Autonomy is arriving faster than many expected, and faster than some are ready for. The prize is clear: AI-led decisioning, real-time data orchestration and end-to-end automated journeys that feel invisible to the customer and dependable to the regulator. The path is equally clear: modernize data and connectivity, re-engineer journeys, operationalize zero-trust and governance and build with partners who can adapt as the technology evolves.

In this landscape, autonomous payments won’t just process transactions, they’ll anticipate needs, strengthen trust and create the intelligent, resilient experiences that define the next era of finance.