Field note2026-09-15T15:35:35.873Z

"Bigger platforms will sort this out for us." Why we don't think that's coming, or coming for everyone.

The search era consolidated around a handful of gatekeepers, so it's a reasonable guess that agentic commerce will too, and that waiting for whichever one wins is the safe move. Genuinely autonomous, personal AI agents don't obviously converge that way, and waiting assumes a consolidation that might just not happen this time.

5 min read

The Selfe team
Agentic commerce infrastructure
Search had a reason to consolidate. Nothing about agents reaching businesses directly obviously needs the same one.

Why search consolidated the way it did

A small electronics retailer decides not to invest in its own agent-readiness, reasoning that Amazon, or whichever assistant ends up dominant, will eventually build the plumbing for everyone and the retailer can just plug in later. It's a familiar bet, it's roughly how search worked out, get found by Google and the rest takes care of itself. It's also a bet on a specific kind of future that isn't guaranteed to repeat.

Search converged on a small number of dominant engines because search itself needed one thing above all: a single index everyone could query, and building that index well was expensive enough that only a few players could do it at scale. That structural reason for consolidation doesn't obviously carry over. A personal agent doesn't need to query one central index of the whole web, it needs to reach the specific businesses relevant to the specific person it's working for, directly.

What a less-consolidated version looks like

If agents end up more personal and more numerous, embedded in different products, configured differently by different people, rather than funnelled through one or two dominant assistants, then there's no single gatekeeper for a retailer to wait on in the first place. Being ready means being reachable by whichever agent shows up, not integrated with whichever platform eventually wins, because there might not be a single winner to integrate with. Waiting for a platform to solve this is its own version of the mistake this hub keeps finding elsewhere: leaning on someone else's interface instead of building the thing that works regardless of whose interface wins.

What's genuinely uncertain here

This isn't a confident claim that no platform will ever dominate agentic commerce. It might. What's overconfident is the opposite claim, that one obviously will, closely enough that a retailer can safely do nothing until it's clear which one. Betting on a business's own facts being connected and reachable works under either version of how this plays out. Betting on a specific future winner only works if that guess turns out right.

Where this shows up for a business

Selfe keeps a retailer's facts reachable by whichever agent comes looking, dominant platform or not, rather than leaving the business waiting on a winner to bet on.

So should we ignore the big platforms entirely?

No, being reachable there too is fine. The mistake is treating one of them as the whole strategy and doing nothing else.

What if a platform does end up dominant?

Being connected and reachable works under that outcome as well. The bet that doesn't work is doing nothing until it's obvious which platform to bet on.