Nobody Buys Everything From One Vendor Anymore

We’ve all seen the robot standoff and dance. Two bots from different manufacturers try all the worst possible ways to get around each other. It’s kinda like the world’s slowest car crash and traffic jam, with bots stacking up in every direction making things worse. Somebody, somewhere, always honks.
Nobody’s fault, usually. The building just never told them who goes first. And here’s the thing: buildings that never told anybody who goes first had both kinds of rules anyway. The official ones in the training binder, and the unofficial ones people worked out aisle by aisle. The buildings that ran well had both.
Seen it a hundred times in the lift world too. Same buildings, same problem. Forklifts, order pickers, walkie stackers and scissor lifts all work the same aisles, and the ones that worked didn’t make every machine identical. They had right of way. Trainees worked their own zones until they earned their way in. The fast stuff stayed fast because the slow stuff knew where to be.
That’s what ISO 21423:2026 gives robots, finally. Not sameness. Just the ability to share a floor without the cheap one gumming up the expensive one.
On November 2, 2026, A3 (Association for Advancing Automation) announced ISO 21423:2026, an international standard developed by ISO Technical Committee 299 and the first non-safety robot standard A3 has helped produce.
In plain terms, it lets robots from different suppliers work the same facility. AGVs and AMRs share what they know: location, speed, direction, condition, what task they’re on, whether they’re available. And not just with each other. With fleet managers, ERP systems, building systems like lifts and automated doors, and the mobile devices people carry.
Think about what that means for a mixed floor. Nobody runs a building where every machine is the same, and you shouldn’t want to. A pallet-mover that does one loop all day doesn’t need the brains of a unit running time-critical routes. What you need is for the simple one to know when to get out of the way, so the fast one isn’t stuck behind it in an aisle, waiting for the slow one to finish being slow.
I’ll be honest, that last part is my read, not the standard’s. The standard doesn’t hand out traffic priority. What it does is let a less capable robot announce what it is and where it’s going, so the faster equipment can plan around it. Like right of way between lift types. The simple bot does its simple job. The expensive one earns its price.
A3’s own claim is more modest than that, and worth taking at its word: fewer robots interfering with each other’s paths, better availability, better efficiency. The standard hands them a common language. The traffic rules are still ours to build.
The old buildings carried the rules in people’s heads, and every shift change dropped a few of them. Now the rules go into the machines. Right of way, written in instead of taped to a column.

Now, a standard on paper isn’t robots cooperating on a floor. Manufacturers still have to build it in, and facilities still have to configure it. This is a direction signal, not a switch you flip Monday morning.
And here’s the part I think gets missed in the announcement-day excitement: today’s version covers the basics. Robots telling each other where they are and what they’re doing. The real prize is what grows out of it, the part where the fleet managers from different vendors and the rest of the building’s systems, the WMS, the ERP, the lifts and doors, stop being strangers and start running the whole floor as one conversation. That’s not what ships this year. That’s what this is for. Think of this standard as the first version of a thing that will keep growing, not the finished product.
It’s not a safety standard either. It doesn’t replace a trained crew, and it doesn’t replace a building that knows its own traffic. Zoning, training tiers, right of way between people and machines, that’s still floor by floor, same as it ever was. What the standard gives you is the language underneath all of it.
What actually changes is growth. Nobody maps out every future need before fixing the worst problem. You start with the job that keeps slowing the floor down, put the right machine on it, and when the next problem shows up you add the next machine. Different vendor if that’s the right vendor. The standard is what keeps the second purchase from becoming a fight with the first.

For years the safe advice was to pick one vendor and stay there, because mixing equipment meant custom integration, separate software, and floors full of machines solving the vendor’s problem instead of yours. A common language loosens that grip. Start with the robot that fixes your worst problem. When the next problem shows up, buy whoever’s machine fixes that one best, and let the standard worry about how they share the aisle.
What I hope comes out of all this is simple: machines that move more seamlessly and fluidly through a building, the way good crews already do. The rules get out of the people’s heads and into the equipment. The binder keeps what it does best, and so does the aisle.
The robots are catching up. About time somebody honked back.



