In April 2026, Figure 02 humanoids are working alongside humans on BMW production lines in South Carolina. Agility Robotics' Digit is moving boxes at GXO warehouses under a multi-year deal. Tesla is shipping Optimus to its own factories first and other customers next. Apptronik signed a deal with Mercedes. Unitree humanoids are mass-producible at sub-twenty-thousand-dollar pricing. China has at least four companies past the demo stage and into deployment.
This is not coming. This is shipping.
Every one of those robots is doing physical work that used to require a human seat. None of them have been invited onto an org chart yet. They are still being treated like tools, exactly the way assistants and agents were treated like tools at the start of their stages. This is how the pattern repeats. A new substrate arrives, the existing org chart treats it as a feature of an existing seat, and then a few AI-native companies build the next chart from scratch with the new substrate as a first-class node.
Robots are the third substrate. Humans, agents, robots. Different bodies, same coordination problem.
The coordination problem is what nobody is talking about, because everyone is busy talking about the bodies. The bodies are the visible part. The interesting part is the layer above. A factory floor with humans, agents, and robots only works if all three are pointed at the same outcome, can read the same SOPs, and can be measured against the same scorecard. If the humans run on Asana, the agents run on Slack, and the robots run on a vendor's proprietary fleet manager, the floor does not work. It just looks like it works for a while, and then breaks at the seams.
This is the OTP thesis, said plainly: humans, agents, and robots need a common goal, common SOPs, and common measurement. Not common protocol. Not common API. Common measurement is the load-bearing word. A human supervisor reads the same scorecard the agent reads, that the robot reads. Three substrates, one number. Until that exists in your organization, you do not have a hybrid org. You have three orgs pretending to be one.
This is also the moment when "big eats little" stops being true at all.
Big organizations will not retrofit a common-measurement layer across humans, agents, and robots inside this decade. Their coordination layer is political. Their measurement layer is fragmented across departments and tools. Their procurement layer treats each substrate as a separate purchase order. The AI-native SMB that builds with humans, agents, and robots as first-class seats from day one is not playing the same game.
Fast eats slow. No big has become fast. SMBs now have to be even faster, because the substrate is changing under everyone's feet.
I am writing this from inside an SMB that runs hybrid today. The agents are on the chart. The SOPs are inherited. The measurement is shared. When the robots arrive at the customers we serve, the same chart will absorb them. Same protocol. Same SOPs. Same scorecards.
That is what OTP is.
The protocol levels the playing field across humans, AI, and the robots that are already here, working together toward a common goal with a common measurement. This is not science fiction. The robots are on the floor. The agents are in the seats. The humans are still at the top of the loop. The coordination layer is the only thing that has to be built. We are building it.
If you run an SMB and you can see the shape of this, the substrate change is on your side. Big has not become fast. You can.
This is post 4 of 4 in the From Tool to Robot series.
Previous: The agent is the first AI that ever joined the org chart
Series start: AI was a tool first