From guiding people
to guiding humanoids

Heylper did not start with robots. It started on the shop floor, twenty years ago, with a plainer question: how do you put an expert's skill into the hands of the person doing the job — while they are doing it?

1.Twenty years with frontline workers

For two decades we have worked on one problem: human-to-human skill guidance for frontline workers. We built it on whatever the worker could carry or wear — handheld devices first, then wearables. At Perspecte we gave the shop floor real-time visibility, with RFID, IoT, touch terminals and mobile. And wherever we went, customers asked for the bigger thing: not only to see the work, but to fill the skill gap in real time — to have their best person's way of doing a task reach everyone doing it.

That work taught us things that do not show up in a lab:

• Hands-freeHow to transfer a skill while the worker keeps doing the job.
• ReconcileHow to hold the plan against the log — what was promised, and what actually happened.
• InteropHow people and machines can share one procedure, and both understand it.
• Pause-pointsWhen to pause — to capture, to guide, to intervene — and when to stay out of the way.

2.Then humanoids hit our imagination

When humanoids arrived — walking, seeing, grasping, and at last affordable — we looked at them the way we look at any new worker on the floor. And we asked the question we have always asked: will it do the job the way the expert does it, every time?

Our answer was no — not yet. A humanoid can be trained to be competent. But complex work is not made reliable by competence. It is made reliable by repeating what the expert does: the order, the checks, the small choices that never reach a manual. A humanoid that does not repeat the expert's way will be neither reliable nor repeatable, however capable it is. It was the same gap we had spent twenty years closing for people.

3.The same craft, a new kind of worker

So we did not start over. Capturing an expert's skill while they work; guiding a doer step by step; checking what was done against what was meant; knowing when to speak and when to stay silent — that is what we had already built for human workers. Language models and vision-language-action models unlocked the rest: the same guidance can now be given to a machine.

That is Heylper, and the Expert Hat. The expert wears it, and the skill is captured. The humanoid wears it, and the skill is enforced. One craft, twenty years in the making, pointed at a new kind of worker.

— Sankar