An AI model that actually understands industrial environments.
Give it a plain-English instruction. It figures out how to do it, adapts when things change, and asks for help when it isn't sure.
X1-D is trained on manufacturing and logistics environments. It can read the scene, understand what you're asking for, and propose how to do it — all before a single safety check runs or a human approves it.
It adapts when the real world doesn't cooperate.
Most automation breaks when something small changes — a bin moved five centimetres, a part arrived in the wrong orientation, a fixture shifted. The machine stops and waits for a human.
X1-D notices the change and works out what to do next.
It reads the scene through cameras, understands the task in plain language, and proposes how to proceed — accounting for the physical limits of the machine and the rules you've set.
If it isn't confident, it stops and tells you why — rather than improvising something unsafe.
Sees the scene
Reads the environment through cameras and understands what it needs to do
Adapts in real time
Adjusts the plan as things change — without stopping and waiting
Three things X1-D does that earlier systems couldn't.
In plain English — what makes it different from traditional industrial automation.
Smooth movements, not jerky commands
Traditional automation gives robots a list of isolated point-to-point commands. X1-D proposes fluid, continuous movements — gentler on hardware, more natural to watch, and easier to inspect before the machine acts.
Handles unexpected changes without guessing
X1-D looks ahead — it considers multiple possible steps before proposing one. If the scene changes mid-task, it re-evaluates rather than improvising. If it can't proceed safely, it stops and escalates instead of guessing.
Works across different robot brands — without reprogramming
Most AI models are trained for one robot type and have to be retrained for another. X1-D separates the intelligence from the hardware — the same model works with KUKA, FANUC, ABB, Universal Robots, and Yaskawa through robot-specific compatibility packs.
Tested in simulation before it touches hardware.
Before X1-D runs on a real machine, it's tested in a virtual version of your environment — where its proposals can be inspected, corrected, and signed off safely.
Simulation runs generate a reviewable record: what the AI proposed, whether it stayed within bounds, where it got blocked, and what needs to change before it's ready for the real robot.
Simulation is evidence that the model is on the right track — not a certificate that it's production-ready. The real machine still requires hardware checks, operator approval, and deployment-specific validation before it runs.
Motion previews
ReplayableEvery proposed movement can be reviewed before the machine acts — shown as a safe preview with the AI's intent and constraints visible.
Safety boundary checks
GatedProposals are checked against your workcell's safety zones and the machine's limits before any live action is considered.
Scene changes
RecoverableWhen the environment shifts — a moved bin, a misaligned part — the model re-evaluates rather than proceeding blindly.
Audit record
CapturedEach simulation run produces a review record showing what the AI proposed, what was blocked, and what needs to change before going live.
Preview before actuation
- The machine doesn't move until every safety check has passed.
- If a proposal is unsafe or unclear, it's blocked, recorded, and sent for human review.
- Simulation results tell you the model is ready for testing — not that it's production certified.
Smart enough to adapt. Safe enough to trust.
In manufacturing and logistics, a mistake is expensive — in downtime, in damaged parts, in liability. An AI that can improvise isn't just helpful — it's a risk.
X1-D proposes movements. Before any movement reaches the machine, the safety layer checks it against your rules. The AI can't skip that check — not by design, not by accident.
Your rules, enforced automatically
If the AI proposes something that violates your operational rules — like tilting an open container — the system blocks it. Business rules are enforced, not just suggested.
Adjusts for what the machine is carrying
When the payload changes — a heavier part, a different grip — the system adjusts how the robot moves, staying within the safe limits you've defined.
Full audit record, automatically
Every decision, every movement, every safety intervention is recorded. If you ever need to prove what happened — for compliance, insurance, or investigation — the record is already there.
Works with your existing setup.
X1-D fits alongside your existing factory control systems. You don't replace anything — you add intelligence to what you already have.
Works with your existing controllers
Connects to standard industrial control systems — PLCs, ROS2, and similar platforms — so you don't need to rebuild your factory floor to use it.
Runs on-site, not in the cloud
X1-D runs at your facility. A network outage doesn't stop your machines — the system continues safely using its last approved setup until connectivity is restored.
Connects to your warehouse and ERP systems
X1-D can read context from your ERP or warehouse management system — so the AI knows what it's supposed to be doing, not just what it's physically seeing.
Start with a technical assessment.
We'll walk through what X1-D can do in your specific environment — what it handles well, where the edges are, and what a deployment would look like.
FAQ
Can X1-D understand plain-language instructions?
Yes. You describe the task in plain English — "pick up the red part and place it in the bin on the right" — and X1-D works out how to do it based on what it can see in the environment.
Does X1-D move the robot directly?
No. X1-D proposes movements — the safety layer checks every proposal against your rules before anything reaches the machine. The AI suggests; a separate system decides whether it's safe to act.
Can I use X1-D with my existing robots?
Yes. X1-D works with KUKA, FANUC, ABB, Universal Robots, and Yaskawa out of the box via Retrofit Studio. For other robot types, an integration engagement is needed.
If the simulation looks good, is the robot ready for production?
Not automatically. Simulation tells you the model is on the right track. The real machine still requires hardware checks, safety validation, operator approval, and deployment-specific testing before it runs.
What happens when X1-D isn't confident about what to do?
It stops and escalates rather than guessing. An uncertain or unclear situation results in the machine pausing and alerting the operator — with a record of what happened and why.
Does X1-D need retraining for every new robot type?
No. The model is separated from hardware-specific behaviour. Robot-specific details are handled through compatibility packs — so the same model can work across different robot brands.