Vision
AI is the foundation. Autonomy and robotics are what it carries next.
We expect the next decade to put autonomous vehicles on delivery routes and humanoid robots in roasteries, factories, repair bays and homes. Because every business here was built on AI from the first line of schema, we can bring those machines into twelve businesses through one system — and put them in the hands of the operators we serve, so that capability scales and the abundance it creates lands with American small businesses.
Only the first phase below is live. Everything after it describes where we believe technology is going and how we intend to use it; the timeframes are our expectation, not a commitment.
Now
AI is the foundation every business runs on.
Orders, pricing, verification and dispatch already run as software with the controls in the database, and every business has its own AI agent answering, diagnosing or advising. This is the layer everything later plugs into.
Orders at a flat 3% across 18 metros, with Basil answering the phone
1.8M+ pros listed, and Zoe answering the phone for field-service businesses
517K+ shops listed, $0 on referrals, and Ray explaining the repair to the driver
Licence-verified marketplace across 13 states — pickup only, by design
Next · 1–3 years
Autonomous delivery, through partners.
Delivery robots and driverless vehicles already carry food in dozens of US cities. We would not build a fleet; we would dispatch to the networks that exist, from the same order, wherever they operate and the law allows. Their economics are still being proven — one more reason to rent before we own.
Autonomous delivery dispatched from the same order — the restaurant keeps its customer instead of renting a driver network that takes a quarter of the ticket
Autonomous same-day delivery, so cold-chain groceries reach further than a driver's shift
Waits on the law. California, for one, requires a licensed delivery employee and prohibits unstaffed vehicles. Where rules change, age and identity checks at the door come first
3–5 years
Robotics on our own floors — humanoids where they pay off.
Automated roasting and robotic machine-tending are proven today. Our operating companies are where we put automation to work first — on our floors, at our risk — and where we expect humanoid robots to take on the varied work in between as they become reliable and economic.
Automated roasting first, then humanoid robots running café service — the same recipe every time, longer hours, and people free to look after the customer
Robotic machine-tending on the domestic zip-tie line from the start, with humanoid robots taking on packing, changeovers and checks as they mature — so American manufacturing competes on cost as well as origin. The line itself is in development today
Within 10 years
Robots in the trades, owned by the operators we serve.
A repair bay or a home is the hardest place for a robot: every job is different. We expect the first humanoid robots there within a decade — remotely supervised at first, more independent as dexterity improves, and carried from job to job by autonomous vehicles. The question that matters is who owns them.
A pro running a crew of people and robots, moved from job site to job site by autonomous vehicles and booked by the same system that answers the phone — the capacity of a national chain, owned by a local business
Humanoid robots assisting with complex repairs in the independent shops we serve — supervised at first, working from the same diagnosis Ray gives the driver, and brought to the car by autonomous vehicle when the job is mobile
Why this order
Easiest and safest first. Hardest and most personal last.
Each phase depends on the one before it, and a machine moves closer to a customer only once it has proven itself further away.
1
Software before hardware
The order, the controls and the dispatch have to exist before a machine can be trusted to act on them. That layer is built.
2
Partners before ownership
Autonomous delivery networks already exist. Dispatching to them costs a fraction of building one and teaches us where it works.
3
Our floors before our customers'
A roastery and a moulding line are ours, structured and repeatable. If a robot is going to make a mistake, it should make it there first.
4
Structured before unstructured
A production line repeats; a repair bay and a home never do. The trades come last because they are genuinely the hardest problem.
The world we are building toward
Capability that scales, and abundance that belongs to Main Street.
Picture a plumbing business with three people and a dozen robots. Autonomous vehicles carry the robots from job site to job site; the same system that answers the phone books the work, dispatches the crew and orders the parts. The owner hires, trains and directs a team that can take on ten times the work — with the capacity of a national chain and the ownership of a local business.
The same pattern holds for the repair shop, the café and the factory line. Technology this capable will create enormous abundance. We intend that it accrue to the American small businesses doing the work, not to the platforms standing between them and their customers.
The evidence
The builders say soon. The sceptics say not yet. We plan for both.
Factory humanoids are already being deployed in pilots, and their builders expect scale within a few years. The hardest part — hands as capable as a mechanic's — may take longer. That is why the trades come last on this page, and why the first robots there will be supervised.
“2027 may see the arrival of robots that can do tasks in the real world.”
“I think Optimus will be the biggest product ever, but it is a very complex problem to solve.”
“Starting in 2028, the Atlas production model will be deployed at Hyundai Motor Group Metaplant America in Georgia.”
“Deployable dexterity will remain pathetic compared to human hands beyond 2036.”
Why we can
Built with AI as the foundation, so new machines plug in once.
Most companies will meet autonomy and robotics one product at a time, each with its own integration. Ours meet them through the engine every business already runs on.
One order model
Every business already turns a request into an order, a dispatch and a record in the same way. An autonomous vehicle or a robot is one more endpoint that receives that order — not a new system to build.
One set of controls
Prices, eligibility, licences and approvals are enforced in the database, not in a prompt. A machine inherits exactly the limits a model does: it cannot set a price, and it cannot release an order the platform has not cleared.
One integration, twelve businesses
Integrate a delivery fleet or a robot platform once and it is available to every business on the engine. That is the difference between AI added to a product and AI as the foundation.
How we will deploy it
Four rules for putting machines to work.
The operator owns the machines and builds the team
Our goal is ten thousand small businesses grown past $10M. Autonomy and robotics are how a ten-person shop competes with a national chain — so in the businesses we serve, the robots belong to the operator, work as part of a team the operator builds, and the productivity gain stays where it was earned.
Law and safety before speed
Cannabis delivery is licensed state by state; autonomous vehicles and robots are regulated city by city. We deploy where it is permitted and safe, and say plainly where it is not.
The same controls, enforced on machines
Age verification, licence checks and order approval live in the platform, so no vehicle or robot can skip them. Constraint matters more as capability grows, not less.
Status stated, not implied
Only the first phase on this page is live. When a pilot starts, it will be labelled as a pilot on the business it belongs to — and not before.
Building toward it
If you build autonomous vehicles, humanoid robots or the platforms behind them and want twelve real businesses to deploy into, we would like to hear from you.