The next twelve months in agentic commerce.
Most writing on where AI agents are heading falls into one of two camps: everything changes overnight, or nothing really changes at all. Neither is honest. Some things about agentic commerce are already settled enough to build for. Others genuinely aren't decided yet, and pretending otherwise doesn't help anyone plan.
8 min read
What's already settled
Say you run a small hotel group, three properties, and spent last year's planning cycle debating which AI shopping platform to build for first: the one built into a general-purpose assistant, a marketplace's own agent, or a specialised travel one. Picking a winner now is a bet on which platform dominates by next year, and that bet might be wrong within months. The groups ready twelve months from now won't be the ones who guessed right. They'll be the ones who got their own facts right once, in a form any of them can read.
An AI agent booking on someone's behalf needs the business's facts to be structured, current and verifiable, whichever platform it's running on. That part isn't in question and hasn't been for a while: every agent, whatever it's built on top of, has to resolve the same handful of things before it recommends or completes a booking, availability, rates, the specific constraints that gate a real request. Building that foundation isn't a bet on any one platform winning. It's true regardless of which one does.
What's still being decided
Whether checkout happens inside the conversation itself or gets handed off to a business's own booking flow, who holds the payment credential and for how long, which of several competing connection standards a given assistant ends up speaking, none of that has settled the way the underlying need for accurate facts has. Anyone claiming certainty here is guessing with confidence, not reporting something known. These specifics are still being negotiated, in public, by the organisations building the standards, and a business doesn't need to have a view on which one wins to do the work that matters regardless.
Infrastructure outlasts any single platform
The distinction that matters for a business deciding where to spend time this year is between betting on a platform and investing in infrastructure. A platform bet is picking ChatGPT's shopping layer, or one marketplace's agent, over the others. An infrastructure investment is getting your own data right, connected and current, in a form that works no matter which agent eventually asks for it. One of those bets can be wrong. The other can't, because it isn't a bet on an outcome, it's groundwork that every outcome needs.
Where this shows up for a business
A business's real booking and pricing systems, connected once through Selfe, are ready for whichever agent shows up, rather than betting on one platform ahead of the others. The rest of this hub goes through specific pieces of what's changing, and what isn't decided yet, in more detail.
Should we wait until the standards settle before doing anything?
No. The groundwork, accurate and connected facts, is needed under every version of how this plays out, so there's nothing to lose by starting now.
Which platform should we build for first?
Picking one platform now is the riskier move. The infrastructure underneath works for all of them.
The businesses that will be ready before anyone asks them to be.
Being ready for AI agents isn't a deadline to hit when the volume arrives. An AI agent already checks the same facts today that it will at scale next year, just less often. The businesses ahead of the curve are doing the unglamorous work now, before anyone's asking them to.
The standards still being written, and why that's not a reason to wait.
The specifications an AI agent uses to connect to a business, prove who it is, and get authorised to spend are all still being actively revised by the organisations building them. That's genuinely unsettled, not a reason to wait. The work that needs doing doesn't depend on any of it finishing first.
The research an agent does before you ever open a search box.
A growing share of product research now happens inside an ongoing conversation with an AI agent someone already trusts, rather than a fresh search each time. Nobody has a precise figure for how much yet, and any business that says otherwise is guessing. The direction, though, is real enough to plan around.
What changes for discovery in the next twelve months.
Not a sweeping prediction. Over the next year, the practical change for discovery is narrower and more useful than "everything will be different": more of the requests reaching a business will already carry real constraints attached, rather than arriving as a vague, browsable query.
What changes for transactions once agents hold payment credentials.
Once an AI agent can hold a payment mandate and complete a purchase on someone's behalf, it's tempting to assume checkout itself transforms. For most businesses, what changes is narrower than that, and what doesn't change is worth stating plainly.
What happens to reviews once agents are the ones reading them.
A star rating tells a person something useful at a glance. It tells an AI agent almost nothing it can act on. As more of a property's reviews get read by agents rather than skimmed by people, what makes a review useful is shifting, and a 4.6 average isn't going to be the part that matters.