Rolling out to the floor
Get Edmund onto one line, confirm the answers are good, then put it in front of every technician who walks past a machine.
Administrator / championOnce your plant is set up, the work shifts from configuration to adoption. The aim is simple: prove Edmund gives good answers on one machine, then widen it without losing that quality.
Pilot, then scale
Start with a single line. Pick a machine your team already asks a lot of questions about, load its documents, connect its data source if there is one, and put a few real questions to Edmund yourself. You’re checking one thing: are the first answers good and grounded, not thin or vague?
When the pilot line answers well, add more machines. Each new project repeats the same pattern — documents in, data source connected if relevant, a quick check that the answers hold up — so you scale on a recipe you’ve already seen work rather than turning everything on at once.
Don’t expand past a line until its answers are good. A thin first answer on day one is what makes a technician decide Edmund isn’t worth opening again. Walk through a technician’s first answer →
The QR + starter-question card
Most technicians will never open this documentation. They meet Edmund at the machine, so put it there. Every project has its own QR code; print it onto a small card and place it at the machine it belongs to.
On the card, print the QR code plus a few starter Suggestions — real questions someone working that machine would actually ask. The first scan then lands on a question, not a blank box, which is the difference between a technician trying Edmund and walking away.
When a machine goes live, tell the technicians who work it that their login is ready and the card is on the machine. They don’t need to wait for anything else to start asking.
Measuring adoption
You can see whether the rollout is landing rather than guessing. Use Analytics and the Daily Chat Activity on the dashboard to watch how often technicians are actually asking questions. Rising activity on a line means the card and the answers are doing their job; a line that stays quiet usually needs better starter questions or more knowledge behind it.
Adoption follows answer quality, so keep each project’s knowledge current. As documents go stale or new machines and procedures arrive, top up the project — thin or out-of-date knowledge shows up first as weak first answers, and weak answers are what stall a rollout. Read the dashboards →
