Agentic Treasury: From Managing Cash to Intelligent Capital

When I was doing some graduate studies at McGill University in the early 2000s, I remember learning about treasury management and the importance of ensuring businesses had sufficient liquidity to meet their financial obligations, manage foreign exchange exposure and make effective decisions about how to deploy capital.

More than two decades later, the fundamentals of treasury management haven't really changed, but the technology available to manage financial operations certainly has.

Despite advances in banking platforms, payments and treasury software, many companies still operate across fragmented financial systems, rely on periodic cash flow forecasts and make important decisions through processes requiring significant human intervention.

What's particularly exciting today is the convergence of artificial intelligence, stablecoins, tokenization and programmable financial infrastructure, which I believe creates an opportunity to fundamentally rethink how businesses manage their financial resources.

We're moving from traditional treasury management toward Agentic Treasury, where intelligent systems can continuously monitor, analyze and eventually execute financial decisions within defined corporate policies.

The implications extend well beyond making treasury operations more efficient.

Consider a multinational business operating across Canada, the United States and Europe, with $20 million in cash spread across multiple financial institutions and currencies, while managing $100 million in annual foreign exchange transactions.

Today, its finance team needs to monitor balances, forecast payments, manage currency exposures and determine where excess liquidity could be deployed, often working across several banking platforms and internal systems.

Now imagine an AI agent continuously monitoring these activities, anticipating payment obligations, identifying currency exposures and evaluating opportunities to improve liquidity or reduce transaction costs.

Rather than waiting for someone to generate a report, the system could proactively recommend actions based on the company's financial objectives and risk parameters.

The CFO and treasury team would continue to establish policies, approve material decisions and oversee risk, but their role would increasingly shift toward supervising intelligent financial systems.

The opportunity isn't about replacing treasury professionals, but enabling them to make better financial decisions with greater speed, consistency and insight.

One area I've been spending more time exploring is foreign exchange and cross-border treasury, particularly as AI creates new possibilities across institutional financial services.

Through executive conversations with Monex Canada, a leader in commercial foreign exchange and cross-border payment solutions, I've gained valuable perspectives on how businesses manage currency exposure, international payments and liquidity.

FX is particularly interesting because even relatively modest improvements in execution costs, forecasting and capital efficiency can generate measurable financial value.

For example, a business managing $100 million in annual FX transactions could potentially save $100,000 through a 10-basis-point improvement in execution costs, while an additional 50 basis points of return on $10 million in available excess liquidity could generate another $50,000 annually.

These are illustrative scenarios, before costs and risk considerations, but they demonstrate why intelligent treasury management could become a compelling opportunity.

Imagine intelligent agents continuously evaluating currency exposures, anticipating liquidity requirements and comparing financial options, helping treasury teams identify opportunities before they become apparent through traditional reporting.

For established providers like Monex Canada, this evolution could create opportunities to build on existing FX and payment expertise with more intelligent, connected treasury services.

This week, Y Combinator highlighted opportunities in its Crypto Request for Startups, including institutional adoption of digital assets, tokenization and the growing role of AI agents in financial services.

What caught my attention wasn't simply the opportunity to build more crypto startups, but the convergence of programmable financial infrastructure with intelligent software capable of interacting directly with financial systems.

Stablecoins are demonstrating how value can move across borders outside traditional banking hours, while tokenized deposits and financial assets could create new ways for businesses to access and manage liquidity.

Meanwhile, AI agents are becoming increasingly capable of evaluating financial information, interacting with service providers and initiating transactions within predefined permissions.

Imagine a corporate treasury agent evaluating cash positions across multiple banks, comparing FX providers and recommending transactions across traditional banking and emerging programmable financial networks.

Of course, a CFO doesn't wake up thinking about how to use stablecoins or blockchain technology, but rather how to reduce costs, manage financial risk and make the company's capital more productive.

The technology only matters if it delivers better business outcomes, particularly when new infrastructure also introduces regulatory, custody and counterparty considerations.

One of the most interesting developments this month is how financial institutions are responding to AI agents attempting to access customer accounts and financial services.

According to Reuters, major institutions including JPMorgan Chase and Morgan Stanley have been restricting unauthorized third-party AI agents, citing concerns about security, customer permissions and the potential for unintended financial transactions.

Meanwhile, fintech companies such as Ramp are taking a different approach, developing financial infrastructure that allows AI agents to interact with expense management, payments and financial workflows through secure connections and defined permissions.

Ramp's approach is particularly interesting because it recognizes that agents need identities, spending limits, approval processes and audit trails, just like human employees.

I don't necessarily see these approaches as contradictory, since banks have legitimate responsibilities to protect customer assets while emerging fintech platforms are demonstrating how authorized agent access could work.

The bigger opportunity is developing financial infrastructure where AI agents can operate securely, transparently and within clearly defined authority.

For financial institutions, this could become an important competitive consideration as customers increasingly expect their AI systems to interact directly with financial products and services.

Much of today's enterprise AI conversation focuses on improving existing operations, increasing productivity and reducing costs, which are important priorities for nearly every business.

However, I believe an equally significant opportunity lies in using AI to create entirely new products, services and businesses.

As I've explored in previous CIO in BETA articles, AI agents are beginning to change how organizations operate, while their growing ability to interact with financial systems introduces new possibilities for how financial services are designed and delivered.

At Highline Beta, we're exploring these AI-driven growth opportunities through our Agentic Finance work, particularly where emerging technologies intersect with meaningful institutional customer needs.

Agentic Treasury is a compelling example because it connects intelligent software with financial infrastructure, established business problems and opportunities to improve capital allocation.

For CIOs, CFOs and corporate leaders, the opportunity isn't simply to automate existing treasury processes, but to rethink how financial decisions are made and how capital can be managed more intelligently across an increasingly connected global economy.

The future of treasury isn't simply moving money faster. It's making capital more intelligent.