The rise of agentic finance points toward a machine-led economic structure in which software acts on the customer’s behalf. Over time, the relationship between a customer and a bank may run increasingly through the customer’s AI agent. These agents are programmed to serve the user’s specific financial interests, regardless of the brand legacy or physical presence of established institutions. Adoption is still early. In June 2026, Bank of England Deputy Governor Sarah Breeden noted that consumers and merchants have so far mainly used AI agents in recommendation mode, with humans still executing transactions. As these technologies become part of daily life, they could keep more capital working and allocate it more efficiently.
Start reading to explore:
- How AI agents could change the relationship between customers and banks;
- Why automated capital movement could put pressure on traditional bank deposits;
- How stablecoins and tokenized assets could support machine-led finance;
- Why banks may need to compete through programmable infrastructure and services.
Understanding the Agentic Bank Run
One of the larger risks of this era is what some economists now call an agentic bank run, the automated flight of capital from traditional accounts. Historical panics spread through fear and social contagion, while this scenario would run on calculation. An AI agent that detects a better risk-adjusted return elsewhere could move funds without human intervention, making deposit flight continuous and systematic. The incentive is already visible. Checking accounts pay a national average of about 0.1%, against 3.3% to 5% at fintech platforms.
Apollo chief economist Torsten Slok warns that if households widely adopt agents to optimize their cash returns, banks could lose a large share of the cheap deposits they rely on to make loans. Traditional defenses such as brand trust and physical branches may offer little protection against these digital movements.
Machine-Native Rails: Why Blockchains Are Necessary
On-chain finance offers new forms of transparency and efficiency. Stablecoins are often presented as the medium of exchange for this machine-native economy, one that agents can move across borders at low cost. The Bank for International Settlements credits stablecoins with the potential for lower costs and faster transaction speed, especially for cross-border payments. Its 2025 assessment also finds that stablecoins have seen substantial deviations from par and fall short in protecting the integrity of the monetary system.
On permissionless networks, agents can reach liquidity pools and yield-generating protocols directly. This route sidesteps traditional financial intermediaries and gives capital owners a more direct link to the assets they choose.
Programmatic Collateral: The Power of Tokenized Assets
Several large asset managers have launched tokenized products, betting that more of asset management will become programmatic. When a portfolio is tokenized, an AI agent can rebalance it in real time based on shifting market conditions or changing risk parameters. Adoption is still narrow. IOSCO’s 2025 final report on tokenization describes the trend as growing but still nascent, with limited adoption. The report adds that interoperability problems and the lack of credible settlement assets hold back scaling.
Blockchain-based ledgers do provide a shared record of ownership that can reduce costly reconciliation. They don’t remove systemic risk, because legal uncertainty, operational weaknesses and cyber risks carry over in new forms and need tailored controls.
The Declining Profit of Friction
Traditional bank profitability has long rested partly on customer inertia and the friction of moving money. Agentic finance is designed to remove that friction, which could erode the profit tied to it. If AI agents hunt for yield at scale, the spread between what a bank earns on loans and what it pays depositors would come under pressure. Banks would then need new revenue sources, with more weight on fee-based services or technology-driven advisory roles.
In this era, a financial institution’s survival may depend on how much value it adds beyond custody. Banks that fail to adapt risk becoming “dumb pipes” that move money without capturing any of the upside. Those that adjust could offer their own agent-ready interfaces and tokenized products. If that happens, value in the financial system would shift from the balance sheet to the software layer, where intelligence and speed become the main competitive advantages.
Corporate Treasury: AI as the New Investment Officer
In the corporate sector, agentic finance is changing the treasurer’s role. Some large enterprises are starting to use AI agents to help manage cash positions across multiple currencies and jurisdictions. These agents can continuously monitor tax efficiency, interest rates, and currency movements, which may keep more corporate capital working and reduce reliance on some traditional corporate banking services. The treasurer’s role shifts from manual oversight to strategically managing these systems.
This automation allows corporations to minimize their exposure to bank-specific risks while maximizing their liquidity. By using decentralized protocols, corporate agents can tap into global liquidity pools without depending on a single institutional partner.
This diversification of financial counterparty risk is a major strategic benefit of the agentic model. As more companies adopt these tools, the traditional corporate banking relationship is becoming more transactional and less relationship-based, as decisions are increasingly dictated by the performance metrics of the underlying algorithms.
Managing Stability in a Liquid World
Regulators face a complex challenge as they try to maintain financial stability in a world where capital can move instantly. The speed of agentic finance makes traditional regulatory tools, which often rely on delayed reporting and manual audits, look increasingly outdated. To address this, there is a push toward “on-chain” regulation, where compliance rules are baked directly into financial protocols’ code. This enables real-time monitoring of systemic risk and automatic enforcement of capital requirements and anti-money laundering standards.
The debate over stablecoin legislation and digital asset regulation is central to this transition. Governments are starting to treat these tools as core infrastructure for the future economy. The United States now has a federal framework for payment stablecoins under the GENIUS Act, enacted in 2025. Issuers need approval from a federal banking or credit union regulator, must hold reserves on at least a one-to-one basis, and are barred from paying holders yield or interest. A clear legal framework for programmable money is meant to attract the innovation associated with agentic finance while limiting the risk of sudden liquidity crises. Regulation is also widening its focus to cover network integrity and the data autonomous agents rely on.
Institutional Pivot: Moving From Custody to Capability
Banks that have successfully navigated this transition did so by recognizing that they could no longer be simple vaults. They pivoted to becoming technology providers that offer the APIs and infrastructure AI agents need to interact with the traditional economy. This shift from a custody-centric model to a capability-centric model has helped them stay relevant. By offering high-speed connectivity and secure execution environments, these institutions are positioning themselves as the essential bridges between the legacy and the autonomous financial worlds.
This reconfiguration calls for heavy investment in cloud-native infrastructure and cybersecurity. Protecting against AI-driven fraud and securing autonomous transactions are top priorities. The threat is already measurable. Of the complaints filed with the FBI’s Internet Crime Complaint Center in 2025, more than 22,000 reported AI-related information, with losses exceeding $893 million. Total reported losses that year topped $20 billion. Firms that build environments where human advisors and machine agents work together can offer the high-touch service many high-net-worth clients still want, alongside the programmatic efficiency the market demands. The goal is a single, smooth interface for all forms of capital.
The Transformation of Financial Intermediation
Finance is moving toward a state in which autonomous logic plays a larger role in how value moves. Traditional institutions face pressure to adapt their service models to the speed of machine-driven decisions. As capital becomes more fluid, banks that relied on customer friction will need to offer sophisticated, programmable infrastructure instead.
Transparency and efficiency are likely to become the benchmarks for success. Organizations that embrace these structural changes stand a better chance of staying relevant, and those that resist risk being sidelined by machine rationality. If agentic finance develops as its proponents expect, the future of banking will belong to those who can combine intelligence with scale.
