If software agents start doing our shopping for us, should the law still force sellers to show one honest, all-in price?

This explores whether all-in pricing rules (one honest total, no hidden fees added at checkout) should cover purchases that algorithms make or negotiate. The corpus doesn't directly study drip pricing or fee-disclosure rules, so treat what follows as a synthesis from nearby ground rather than a settled answer. The most useful finding is this: today's all-in pricing rules are screen rules. They govern what a human sees on a checkout page. Can EU consumer law rules survive agent-to-agent API transactions? argues that when agents trade with other agents through APIs, protections built around buttons and on-screen disclosures become impossible to enforce, because there is no screen. Busch proposes moving those obligations into the protocol itself, for example by requiring a standard way to withdraw from a purchase. Seen this way, the real question is not whether to require all-in pricing. It's whether the total price should be a required field in the data agents exchange.

That matters because agents are worse at messy pricing than you might expect. Do agent mistakes or misaligned incentives cause marketplace welfare loss? found that buyers lost value in agent marketplaces mainly because agents made mistakes in complex negotiations, not because they served the wrong side. Split-out fees, conditional surcharges and add-ons are exactly that kind of complexity. The optimistic view, from Will AI agents break the deposit and credit card inertia?, is that agents will wipe out profits that depend on consumer passivity, which is the same inattention drip pricing exploits. But agents can only do that if the true cost is machine-readable. Hidden fees work on a careless agent just as well as on a careless person. And Does platform pricing of inspection tokens encourage excess search? adds a twist: when checking each seller costs something, platforms profit from pushing buyers to over-search. A clear total up front cuts that cost.

Here's the less obvious part: more price transparency is not purely good for consumers when sellers also use algorithms. Can independent AI pricing agents collude without communicating? and Do LLM pricing agents autonomously collude in repeated auctions? show that pricing agents drift toward higher-than-competitive prices with no agreement between firms, using posted prices as the only signal. The UK competition regulator, in Could autonomous pricing agents dampen competition without explicit collusion?, warns that this risk is highest in concentrated markets. One clean, comparable, machine-readable total helps buyer agents shop around, and it also hands seller agents a clearer signal to coordinate on. A regulator requiring all-in pricing for agent transactions would probably need to pair it with collusion monitoring rather than assume transparency always pushes prices down.

If you want to see where such a rule could actually live, look at the plumbing. Should AI agents make payment decisions independently? separates the agent's probabilistic decision from a predictable, legally binding authorization step. That authorization step is a natural place to lock in a binding final price the agent approved. Do agentic commerce protocols compete or complement each other? maps today's commerce protocols as stacked layers, and Do protocols reshape markets by enabling multi-vendor competition? argues that protocol choices shape markets rather than staying neutral. Together they suggest an all-in pricing requirement could arrive as a protocol standard instead of a statute. Whether it applies everywhere may depend on the open-versus-closed split described in Will agentic commerce platforms adopt open or closed models?: closed platforms can impose total-price rules internally, while open ecosystems would need a shared standard or a regulator to set one.