The emergence of autonomous treasury has spurred an intense competitive drive within the corporate finance sector, accompanied by ambitious industry benchmarks. However, companies are expected to adopt these changes at varying paces.
This transition to an autonomous treasury is significantly altering the landscape of corporate finance, fueled by innovative strategies and technologies—ranging from self-healing cash forecasts to AI-powered liquidity solutions—that are replacing outdated systems and enhancing yield.
To unlock the full advantages of autonomous treasury, finance leaders are making targeted investments in critical areas that will hasten this shift. Sayantan Chakraborty, head of Digital Payments at Fiserv, emphasizes that today’s treasurers possess ample visibility but lack the tools necessary for real-time action. “The gap isn’t analytics; it’s execution,” he notes.
While agentic AI is capable of predicting cash flow and generating funding directives, Chakraborty highlights that much of the existing corporate infrastructure operates in batch mode. He identifies three key components necessary for improvement: first, comprehensive real-time cash positioning; second, a rule-based system for timely fund transfers across both instant and traditional payment channels; and third, the integration of features such as tokenized deposits and programmable payments.
Despite technological advancements, the process still relies heavily on human expertise. Chakraborty advises improving existing ERP systems rather than waiting for complete overhauls. “Consider it an AI-powered autopilot added to an older cockpit,” he suggests. “Regulations are followed, actions are taken, and audit trails are maintained without necessitating an all-at-once overhaul, all under the supervision of trained personnel.”
Chakraborty criticizes the outdated notion of extensive, multi-year upgrades, advocating instead for the adoption of an agile, 24/7 automation framework that manages real-time balances, regulations, and payments.
As real-time reporting and instant payment systems gain traction, he anticipates that the current approach of pre-funding account balances prior to deadlines will soon become obsolete. Instead, “agentic AI will shift treasury operations from once-daily instructions to ongoing, on-demand funding whenever execution aligns with intent across all channels.”
This transformation is likely to reduce the idle-balance float and prompt banks to concentrate their earnings on services like 24/7 clearing, intraday credit, and real-time liquidity.
In late 2023, Siemens adopted J.P. Morgan’s programmable payment technology (previously known as Onyx, now Kinexys) to enhance its autonomous treasury capabilities. By implementing advanced programmable payments via JPM Coin, Siemens automated cash management by executing transactions based on predefined criteria. This strategy tackles the issue of inefficient pre-funded balances, ensuring that funds are only allocated to accounts when payments are imminent. Additionally, when account balances fall beneath a specified level, the system automatically draws funds from a central cash reservoir, allowing Siemens to maintain near-zero balances in individual accounts.
According to Heiko Nix, Siemens’ global head of Cash Management and Payments, the primary hurdle is not technological but rather a cultural shift within finance and treasury departments. “For nearly every technical challenge, a solution exists. However, transforming long-standing processes and changing perceptions of treasury’s role requires significantly more time and effort,” he explains. He advises that change does not need to occur simultaneously across the organization; instead, creating sufficient momentum is crucial for meaningful transformation.
A New Approach to Control
John Stevens, senior vice president and global head of Capital Markets, Financial Institutions & Working Capital at Kyriba, argues that AI presents a strategic opportunity. “AI has the potential to transition working capital management from a historical reporting function to a forward-looking control tower,” he states. By shifting focus from past occurrences to real-time future optimization, tasks that once required prolonged manual effort can now be completed instantaneously, facilitating more timely and informed decision-making.
Stevens warns that organizations must collaborate closely with technology providers to ensure the safe implementation of AI. He notes that there will not be a single autonomous product that meets every treasury need. Instead, the future will be characterized by a “composable” approach, which he stresses must be precisely defined.
While Kyriba App Studio offers a layer for custom integrations and workflows, Stevens clarifies that it is not a toolkit for building AI agents. The agentic AI layer, referred to as TAI, is developed by Kyriba and maintains a “human in the loop” philosophy.
Utilizing a third-party model does not inherently detract from an AI tool’s intelligence, nor does strictly in-house modeling guarantee superior performance, Stevens argues. “In the treasury context, the key factor is whether the AI can operate safely and consistently under regulation,” he asserts. TAI is not intended to exclude external large language models; rather, it uses a prominent external model (Anthropic’s Claude) in a controlled deployment, emphasizing strict limitations on data access and an audit trail of actions.
This structured approach enables the AI to assist in generating insights, summaries, and alerts, while all matters affecting payments, liquidity, or risk remain under rigorous platform governance, approvals, and policy-driven frameworks.
Transforming Corporate Finance
The anticipated benefits of autonomous treasury have sparked a competitive rush towards autonomy, accompanied by aggressive industry goals and a race towards fully automated platforms.
HighRadius has recently upgraded its agentic AI platform, aiming for over 90% automation for the Office of the CFO by 2027. This initiative will deploy AI agents across six product suites and 20 products, covering accounts receivable, payables, treasury, and consolidation. The introduction of 186 agentic AI agents, announced earlier this year, aligns with HighRadius’s vision of achieving a fully autonomous platform, with cash application and forecasting already showing 90% automation without human intervention.
CEO Sashi Narahari regards agentic AI as a stepping stone towards achieving full autonomy across all products, which is defined as over 90% touchless end-to-end processes by 2027. He emphasizes the importance of this objective, noting that failure to achieve it would threaten the survival of the company.
What of mid-tier banks that may not wish to undertake a sweeping transformation? Chakraborty suggests that a single, dependable orchestration endpoint is preferable to multiple disconnected APIs. “A real-time balance and payment execution API is essential, providing visibility into positions, limits, and instantaneous transfers through a single, robust interface,” he says. Such integration, especially when coupled with tokenized deposit movements, can be advantageous wherever feasible.
Ultimately, the journey towards autonomous treasury—led by innovators like Siemens and propelled by advancements in agentic AI—is fundamentally redefining corporate finance. This transition is not merely focused on incremental efficiency gains but is recognized as a strategic necessity to optimize yield, secure real-time liquidity, and transcend the limitations of legacy systems. Corporate treasurers embracing this transformation are offered a tactical path to a future-ready role, while financial institutions are urged to align their services with the demands of this new era defined by continuous, intelligent, and timely financial governance.