Agreeing a pilot scorecard you can both live with
Five lines, two arguments, and how to settle them before integration starts.
Case studies
Scorecards,written beforethe pilot starts.
In brief
The studies below are modelled pilot scenarios, not customer results. They show exactly how a Weldeon pilot is scoped, measured and reviewed — so you know what you would be signing up to.
Next step
A note on evidence
Why there are no customer logos on this page
Weldeon is early. Customer, revenue and traction status is not public, and we are recruiting three to five design partners for paid pilots now.
We could dress modelled scenarios as case studies. Fabrication buyers see through that immediately, and it would poison the one thing this category actually runs on — an honest scorecard.
So the studies below are labelled for what they are: worked scenarios built from public process economics and our own pilot methodology. When real scorecards are signed and cleared for publication, they will replace these, and we will say which is which.
Scenario 01
Structural steel — connection welding line
[SCENARIO] A structural fabricator running a robotic connection line under AWS D1.1, losing rate to fit-up variability and special-inspection rework.
01
Fit-up first, then the arc
The pilot scope is one line, two agents: fit-up and tack, then weld and parameter. Baseline is four weeks of measured first-pass yield and rework hours. The scorecard targets a five-point FPY gain and a 20% reduction in rework hours over ten weeks of supervised autonomy.
- Scope: one connection line, D1.1
- Targets: +5 pts FPY, −20% rework hours
- Review: week 12, on the floor, against the agreed baseline
Scenario 02
Shipbuilding — panel line distortion
[SCENARIO] A yard whose block erection schedule is set by the straightening bay, not the welding bay.
02
Plan the heat, save the block
The scope is the panel line and the distortion-and-stress agent. Sequence and heat balance are planned on the twin; pre-camber and restraint are recommended per panel. The scorecard targets a 25% reduction in straightening hours and out-of-plane distortion inside 3 mm on 90% of panels.
- Scope: one panel line
- Targets: −25% straightening hours, 90% inside 3 mm
- Measured with the yard's existing laser survey
Scenario 03
Pressure vessels — the NDT bay
[SCENARIO] An ASME BPVC shop where re-inspection has become the constraint and a missed indication is a safety event.
03
Read first, weld later
The scope is inspection only: digital radiography and phased-array ultrasonic interpretation running in parallel with the shop's qualified inspectors for eight weeks, with no write authority anywhere. The scorecard measures agreement rate, missed-indication rate and inspector hours recovered.
- Scope: NDT bay, advisory only
- Targets: agreement rate and inspector hours
- No autonomy granted until agreement is proven
Scorecard
The four lines every pilot is judged on
FPY
first-pass yield against a measured baseline period
Rework
rework and straightening hours in the pilot scope
Escapes
defect escapes to re-inspection or the field
Takt
station cycle time against the line's takt
Figures on this page describe the pilot scorecard Weldeon underwrites with design partners. They are target outcomes, not audited customer results; customer names and traction are not yet public.
Method
How a pilot is run
The same five phases whichever agent leads. The point is that nothing about the measurement is decided after the result is known.
- 01
Scope
One cell or line, one or two agents, chosen where the pain already has a number attached to it.
- 02
Baseline
Two to four weeks of measurement using the shop's own instruments and definitions, agreed in writing.
- 03
Shadow
Weldeon runs read-only, logging the decisions it would have made, until agreement is convincing.
- 04
Bounded autonomy
Write authority granted inside the qualified envelope, with weekly floor reviews and a stop policy.
- 05
Review
Scorecard reviewed against the agreed targets, in the open, including the lines we missed.
“Publish the misses as well as the hits. A fabricator can smell a curated result from the car park.”
Measurement
How each line is actually measured
The definitions that prevent an argument in week twelve.
- First-pass yield
- Welds accepted at first inspection ÷ welds made, counted at the same inspection point as the baseline.
- Rework hours
- Booked repair, grinding and straightening hours attributed to the pilot scope, from the shop's own time system.
- Defect escapes
- Indications found at re-inspection or downstream that were not called in the pilot scope.
- Takt
- Station cycle time measured from the shop's existing production system, not from Weldeon telemetry.
- Baseline integrity
- Baseline period, product mix and shift pattern fixed before the pilot begins; mix shifts are disclosed in the review.
Related
How we think about proof
Longer-form writing on measurement, attribution and pilot design.
Attribution: the rework hour and the station that caused it
Why rework is almost always booked to the wrong station, and what that hides.
What a good baseline period looks like
Mix, shift pattern, seasonality — and the four weeks that decide whether your pilot means anything.
Questions
Case study questions
-
No. Every study on this page is a modelled scenario, labelled [SCENARIO]. Real, cleared customer scorecards will replace them, and we will mark clearly which is which.
-
Yes. The scenario assumptions, process economics and scorecard template are available on request and are the same artefacts we use to scope a real pilot.
-
Then we say so in the review, in writing, including which of our assumptions were wrong. A failed pilot that is honestly reported is worth more to us than a converted one that was fudged.
Begin
Bring autonomy to thearc, the fit-up andthe inspection bay.
Book a pilot on one cell or line. We agree the scorecard before we start — first-pass yield, rework hours, defect escapes, takt — and we report against it in the open.