A refund sounds like a simple rule: if the booking is refundable and costs less than $200, approve it. A smart contract can enforce that rule. But first, someone has to tell it whether the booking actually qualifies. The contract can't open the booking page, read the policy and interpret what it means on its own. That gap matters. You can put the payment on-chain, but if the decision depends on a single outside report, part of the trust still sits off-chain. @GenLayer's Intelligent Contracts can fetch the live booking page and ask a language model whether the booking is refundable and under $200. Here's the part I find most important: one model doesn't get the final say by itself. A validator proposes an answer, while other validators run the contract independently with their own models. If they agree, the result is accepted. If they disagree, another leader gets a chance to try again, and anyone can appeal for further review. That doesn't guarantee every judgment is correct. But it gives the system a way to handle disagreement instead of blindly trusting one answer. Smart contracts follow predefined rules. Intelligent Contracts can evaluate information those rules alone can't interpret. Would you trust this process to decide a refund, or would you still want a human review?