Fabrication, assembly, food, print

When the ERP and the floor disagree,the floor is right.

Every plant we walk has two versions of its process: the one in the system and the one that gets product out of the door. The gap between them is not a compliance problem. It is the map of where the work actually is.

Three days on site · three days of analysis · the report is yours either way
12

Open source projects

Public on github.com/JIGGAI

4

AI plugins

Published on npm, free to install

65.4K

AI plugin installs

65,377 to date, per npm

2.8K

AI monthly plugin installs

2,811 in August 2026, per npm

Sound familiar

What we hear in the first hour.

Not a generic list. These are the ones that come up in businesses shaped like yours.

01Scheduling

Nobody trusts the system, so the schedule lives on a whiteboard.

The whiteboard is the real plan. Whatever replaces it has to be as fast to change as a marker, or it will not be used either.

02Quoting

Quoting takes three days because one person has to price it.

Usually the highest-value automation on site. The estimating logic exists — it is in somebody’s head and in the last two hundred quotes.

03Materials

We find out about a shortage when the line stops.

The data to see it coming is already in the system. Nothing is watching it, because watching is a job nobody was given.

04Handover

Shift handover is a conversation, and things get lost in it.

Handover notes are the cheapest possible finding: high frequency, low judgement, and everybody already agrees the current version is bad.

05Quality

Certificates and traceability paperwork take longer than the job.

Compiling evidence from systems that already hold it. Assembly, not authorship.

The rule

The floor is not resisting the system. It is routing around something that does not fit the work.

Which is why we start on the floor rather than in the ERP. A recommendation that contradicts how the job is actually done gets ignored, and rightly.

What we expect to find

The software already in the building.

We do not arrive with a platform to sell you. We arrive expecting these, and the work is usually in the gaps between them.

01

The ERP or MRP

Epicor, Global Shop, SAP Business One, Sage X3. Correct in principle, worked around in practice.

02

The scheduling whiteboard

Or the spreadsheet that photographs it. The actual plan, and the first thing we ask to see.

03

Shop-floor data capture

Barcode scanners, tablets, or paper travellers. Where the gap between recorded and real usually opens.

04

Quoting and estimating

Nearly always a spreadsheet with one author, and nearly always the highest-value thing in the building.

05

Quality and traceability

Certificates, batch records, non-conformance logs. Evidence assembly, repeated by hand.

06

Maintenance records

A CMMS if you are lucky, a shared drive if not. Where unplanned downtime was usually predictable in hindsight.

A worked example

Getting the quote back while it still wins the job.

Roughly what one entry in your report looks like — a real shape, with the numbers changed.

What we saw
Enquiries arriving by email and phone, then queuing for the one estimator who could price them. Material costs were looked up manually per quote, and similar jobs already quoted were found by memory rather than search.
What it costs today
A turnaround slow enough to lose work, in a market where the first credible quote often wins the job. The estimator spent more time on lookups than on the judgement only they could apply.
What we would build
A drafting step that matches an incoming enquiry against previous jobs, pulls current material costs, and produces a priced draft with its assumptions listed. The estimator adjusts and sends. Every override is recorded, so the matching improves against real decisions rather than a guess.
How you would know it worked
Time from enquiry to quote sent, win rate on quotes returned quickly, and how often the estimator overrides the draft. A high override rate means the matching is wrong and we fix that rather than adding steps.
And what we would not automate

Sending the quote. Pricing is commercial judgement about a customer, a relationship and how badly you want the work this month — and a system that sent them automatically would eventually send one you would not have. The estimator stays the last step.

The arithmetic

What coordination work costs you.

Three numbers you already half-know. Move them until they look like your business.

12

re-keying, chasing status, producing the same document again

6h

on that work specifically, not their whole job

$65

salary, tax, benefits, desk

Coordination work, per year
3,312hours
Plausibly retired by agentic automation
2,484hours

A 75% capture rate. The other twenty-five per cent is judgement, exceptions, and not wanting to look greedy.

1.4
FTE equivalent
$161,460
Per year

This is arithmetic, not a finding. It rests on three numbers you guessed. The assessment replaces all three with numbers we observed — and tells you which of those hours are actually worth automating.

See what a real finding looks like →
What happens

What the six days look like.

The most common question we get is not about AI. It is what these people will actually do in my building.

Assessment · 3 days on site · 3 days analysisNo prep required
  1. Day 1Walk the floorWhoever is on shift

    We start where the work happens, not in a meeting room. Nobody prepares anything, and the first day is mostly watching.

  2. Day 2Sit with the people doing itOps, admin, finance

    Conversations with the roles that touch the work most, and a real task followed end to end — including the parts that happen in a group chat.

  3. Day 3Systems, then a read-backWhoever holds the logins

    What you run, what talks to what, and where a person is currently the integration. We tell you what we saw before we leave, while it is still cheap to correct.

  4. Days 4–6Research and discoveryOur desks, not yours

    Away from your building. We cost the work we watched, model the alternatives, and test the shortlist against your own numbers rather than a framework.

  5. +1 weekOne recommendationPresented in person

    The single change worth making first, specified precisely enough to build — with the ranked analysis behind it and the list of what we would not automate.

Before you ask

The questions this raises.

Do you need to stop the line?

No. We walk it while it runs, which is the only way to see the workarounds. We need someone who can explain what we are looking at, and a slot with whoever holds the system logins.

Our ERP vendor says they already do this.

Sometimes they do, and the finding is that you are paying for a module nobody switched on. That is a good outcome and it goes in the report as one.

The floor is sceptical of head-office projects.

Reasonably. That is why the interviews are with the people doing the work rather than about them, and why the report names what should stay manual as prominently as what should not.

We are seasonal. Does timing matter?

Visit during a busy period if you can. A quiet week shows us a version of your operation that does not have the problems worth solving.

Start with half an hour.

We'll tell you whether an assessment would pay for itself in an operation like yours. Sometimes the answer is no.