Centralised Exchanges Must Adapt to the Rise of AI Agents

By Ben Caselin, CMO, VALR

Originally published by Crypto Citizens Network.

AI agents are moving from experimentation to execution in financial markets, and centralised exchanges will remain critical to this shift, but this will require a reconsideration of how identity, compliance, and accountability work for non-human actors.

The early wave of “AI in crypto” was mostly noise. Tokens with vague claims and little substance dominated the narrative. That phase is ending. What’s emerging now is more practical and more consequential: autonomous systems that can analyse data, allocate capital, and execute transactions with minimal human input.

This shift is no longer theoretical. AI-driven trading strategies, API-based automation, and early agent frameworks are already interacting with crypto infrastructure. In doing so, they are exposing a fundamental gap between systems designed for human users and a new class of machine participants.

The Rise of Agentic Finance

AI agents, systems that can perceive, reason, and act independently, are beginning to manage portfolios, execute trades, and coordinate on-chain activity. This is giving rise to what can be described as “agentic finance,” a model in which economic activity is increasingly carried out by software rather than people.

Blockchain infrastructure is naturally suited to this shift. Programmable money, near-instant settlement, and open APIs make it easier for autonomous systems to participate directly in markets. Traditional financial systems, built around human intermediaries and slower settlement layers, are less adaptable.

The result is a new layer of economic activity that is already forming at the edges of the market and is likely to expand quickly.

Where agents will operate

New technologies tend to gravitate toward environments that accommodate them with the least friction. In this case, that points to open, decentralized infrastructure.

Permissionless access, composability, and minimal onboarding requirements make decentralised platforms an obvious starting point for autonomous agents. Early experimentation, particularly in on-chain trading and micro-economies, is already happening in these environments.

But this does not mean centralised exchanges are at a disadvantage. For agents operating at scale, they may become essential infrastructure.

Deep liquidity, fast execution, robust risk controls, and secure custody are not optional for systems managing meaningful capital. Regulated platforms also provide accountability and trust signals that matter not only to institutions, but to the humans and entities ultimately responsible for deploying these agents.

The question is no longer whether AI agents will participate in markets, but where they will be able to.

The identity problem exchanges must face

The real challenge for centralised exchanges is how to reconcile autonomous agents with frameworks built around human identity and legal accountability.

What does onboarding look like for a non-human actor? Traditional KYC models are not designed for software. New approaches may involve cryptographic credentials, verifiable agent identities, API-based permissioning tied to legal entities, or reputation systems based on audited code and historical behaviour.

Technical safeguards are relatively straightforward. Secure authentication, rate limiting, real-time monitoring, and standardised interfaces already exist. Many exchanges support API-based trading that allows automated systems to operate within controlled parameters.

The deeper issue is structural. Responsibility must still rest with identifiable individuals or entities. Ensuring that an agent’s behavior remains aligned with platform rules, preventing manipulation, and establishing trust in its provenance are not purely technical problems.

They require new models of accountability that bridge software autonomy and regulatory expectations.

Early signs of adaptation

Incremental progress is already visible.

API infrastructure enables automated trading and portfolio management at scale. In-platform AI tools are improving user interaction and decision-making. Some exchanges are beginning to explore how autonomous systems can operate within regulated environments, even if full independence is not yet possible.

These are early steps. They show that the gap between human-centric systems and machine participants is being actively explored.

Adaptation is not optional

The trajectory is clear. Agentic finance will expand as AI systems become more capable and more widely deployed.

If centralised exchanges adapt, they can anchor this new layer of economic activity by providing the liquidity, security, and regulatory credibility that autonomous systems will need as they scale. They can become the interface between machine-driven strategies and institutional-grade markets.

If they do not, the outcome is equally clear. Liquidity, innovation, and increasingly sophisticated agent-driven strategies will migrate toward open, on-chain environments that are better suited to autonomous participation.

The opportunity is significant, but so is the risk of being bypassed.

Centralised exchanges have long been the gateways to crypto markets. In an agentic financial system, they will need to evolve into something more: infrastructure designed not just for humans, but for machines acting on their behalf.

About the Author

Ben Caselin is Chief Marketing Officer at VALR.com, the Pantera-backed crypto exchange. Headquartered in Johannesburg and the largest exchange by trade volume on the African continent, VALR serves over 2,000 corporate and professional investors and more than 1.8 million traders worldwide. Drawing on years of experience in the digital asset space, mostly in Hong Kong, the UAE and now South Africa, Ben focuses primarily on driving the adoption of bitcoin in emerging markets and advocates for an approach to innovation on the basis of spiritual principles.

Risk Disclosure

Trading or investing in crypto assets is risky and may result in the loss of capital as the value may fluctuate. VALR (Pty) Ltd is a licensed financial services provider (FSP #53308).

Disclaimer: Views expressed in this article are the personal views of the author and should not form the basis for making investment decisions, nor be construed as a recommendation or advice to engage in investment transactions.