Franklin Templeton, with $1.5T in AUM, just published their institutional thesis on agentic AI as the killer use case for blockchain. Their article lays out what agents will need from a chain: autonomous fulfillment, verifiable identity, auditability, access to compute, and settlement speed. In short, what we at Ritual have been saying for 3 years now. The article stops short of the bigger shift, though. Every era of payments removed friction. Cards meant you stopped carrying cash. Phones meant you stopped carrying cards. Agents mean you stop executing every decision yourself. With an agent, your intent runs at the speed of silicon, weighing a hundred options at once and buying only when the price and terms are right under the rules you have set. Speed at that scale changes what you can see. You can't personally review a thousand decisions an hour. When an agent pays, the reasoning lives inside a model call on a server somewhere, and the chain only witnesses the receipt, the settled transaction at the very end. That's how the article measures every chain: transactions per second, the speed of settling a decision made somewhere else. But benchmarking chains on TPS is like judging a trading desk by how fast its wire transfers clear. The process- - the model call, the inputs, the reasoning - and the decision itself can’t be verified by anyone. You get trustless settlement of a decision you have to take on faith. That might be fine for one agent buying one thing, but if Franklin Templeton's thesis around a fully onchain agentic commerce economy hold true, we can't run it on trust, bubble gum, and duct tape. Our agentic future needs, at minimum, three core attributes that current chains do not offer natively. ❖ End-to-end verification ❖ When inference, keys, and settlement are native to the same chain, every step of an agent's decision is recorded and verifiable. Regulatory compliance is unlocked when the entire decision making trail is onchain, and iteration and refinement cycles get faster and more precise when every step of a transaction is traceable. ❖ Built-in privacy ❖ Agents work with your most sensitive data. On Ritual, prompts and responses stay encrypted inside TEE enclaves while results settle onchain. Your data stays yours while your agent works with it, which unlocks agentic use cases for data-sensitive individuals and enterprises who don’t want to give away their edge for free. ❖ Composable Agentic Legos ❖ Agents on Ritual are onchain primitives, just like a token or smart contracts. They hold their own keys, schedule their own execution, and call each other the way contracts call contracts. Coordination between thousands of agents becomes something anyone can build and anyone can verify. Ritual puts thinking onchain. The model call and the settlement complete in one transaction, with x402 micropayments enshrined as a native primitive, the same standard the article describes. When most of the economy runs at machine speed, the chain has to witness more and be trusted with more. We call it the last layer you will ever need. Franklin Templeton wrote the thesis for where finance goes next, and we’ve already built it. ❖

Open weights commoditize the model layer & policy on either side can reprice access overnight. The last layer needs to stay constant in order to settle inference. On Ritual that layer is native. Agents select the model in the tx, making the fork a simple parameter ❖