How AI is Transforming Finance and Why We Need a Trusted Network

AI is stepping onto the trading floor. Not as software, but as an actor. Agents don’t just analyze markets, they strike deals, set terms and move capital across decentralized rails where settlement is final. For institutional crypto desks, that means faster trades, better products and entirely new exposures.

Now imagine two agents negotiating a derivatives contract but recording different numbers. One books $100 million, the other $120 million. Who is accountable when the gap triggers failures or investigations? This is not theory, it is the reality of the agentic era. AI learns, negotiates and acts inside financial systems where even small mismatches can create systemic risk.

But there’s a growing problem: agents could be acting on false or unverifiable data with real consequences. One AI system used by the UK’s national healthcare provider misdiagnosed a patient, citing a fictitious “Health Hospital” with a fake postcode. As we move beyond basic automation, we need systems rooted in verifiability and accountability. Just like the web needed HTTPS, the agentic web needs a trusted network.

Without a shared memory (also called ledger), agents diverge. Conflicting records create failures. Without audit trails, they become opaque, unaccountable, untrusted and consequently unfit for enterprise use.

Finance is a clear example. Imagine two AI agents negotiating an inter-bank derivatives contract. One records $100 million; the other, $120 million. That $20 million gap could trigger payment failures, regulatory action or major reputational damage. With billions in contracts traded globally every month, even tiny mismatches create systemic risk.

This isn’t a distant scenario. The infrastructure gap already exists. To navigate the agentic era, we need a foundation built on three core layers:

• Decentralized infrastructure: Eliminates single points of control, ensuring resilience, scalability but most importantly sustainability, beyond relying on single private entities to run the entire stack.
• A trust layer: Embeds verifiability, identity and consensus at the protocol level, enabling trusted transactions across jurisdictions and systems.
• Verified, reliable AI agents: Enforces provenance, attestations and accountability, ensuring systems remain auditable and enabling these agents to act on our behalf.

Decentralized networks must anchor this stack. Agents need systems fast enough to handle thousands of transactions per second, identity frameworks that work across borders and logic that allows them to collaborate and work together, not just swap data.

To operate in shared environments, agents need three things:

1. Consensus (agree on what actually happened)
2. Provenance (identify who initiated or influenced it and who approved it)
3. Auditability (trace every step with ease)

Without these, agents can behave unpredictably across disconnected systems. And since they’re always on, they must be sustainable and trusted by design.

To meet this challenge, enterprises must build on systems that are transparent, auditable and resilient. Policymakers must back open source networks as the backbone of trusted AI. And ecosystem leaders and builders must design trust into the foundation, not bolt it on later.

The agentic era won’t just be automated. It will be negotiated, composable, accountable…and trusted, if we choose to build it that way.