The plain list a plant manager works through to know whether the floor can support AI yet. No score to chase, no jargon. Just the questions to answer honestly about your own plant, in the order they actually have to happen.
Read this first · How to use this
This is a checklist you work through about your own plant, not a quiz about ours. Every item is a plain yes or no you can answer from the floor. It is built around one idea: AI runs on data a system can read, and most of a high-production plant still runs on paper, one person's spreadsheet, and machines that talk to nothing. You cannot point AI at a record that only exists as handwriting.
There are no invented numbers on this page. The only figure here is a public rule with a link. Everything else is something you check against your own floor, because an average across other people's plants would not tell you what to fix on yours.
Why this comes before buying AI
Manufacturing software is sold as though the data already exists. The pitch assumes records sit in systems, machines report their numbers, and the figures are current. On most floors that is not the case: much of the week starts as handwriting at a station, and the software only meets it after somebody re-types it.
That gap is the whole problem, because an agent that cannot read the record cannot act on the record. A copilot answering from a system that holds a third of the week gives confident answers about a third of the plant. So the honest first question is not which AI to buy. It is how much of your floor an AI could even see today. This checklist walks you to that answer.
The order it has to happen in
The plants that are actually ready did the same unglamorous thing in the same order: digitize, connect, unify, before any AI. Digitize gets records off paper at the station. Connect gets machines and systems sharing what they already know. Unify pulls all of it into one live layer that is current and entered once. Only then does AI have something to stand on. That sequence is why the checklist is grouped the way it is, and it maps onto the five things a paper audit measures: paper capture points, data latency, machine data escape, re-keying, and trace time.
The five checklist sections
Here is the shape of it. Each section is one part of the sequence, tied to one thing the audit measures. The full checkable items under each are below, sent to your work email or revealed here on this page.
Section 01
Digitize: how much still starts on paper
Measures: paper capture points
Every place a person writes something down that later has to become data. Batch sheets, quality checks, travelers, rounds, downtime logs, handoffs.
Section 02
Connect: what your machines and systems actually share
Measures: machine data escape
Whether counts, states, faults and cycle times leave the machine at all, and whether your ERP, MES, quality and warehouse systems talk without a person carrying numbers between them.
Section 03
Unify: whether your numbers are current and entered once
Measures: data latency and re-keying
How old a number is by the time someone acts on it, and how many hours a week your team spends typing figures that already exist somewhere else.
Section 04
Prove it: can you trace one real lot fast
Measures: trace time
The single most honest test of whether records are connected, because a real trace has to cross receiving, production, quality and shipping in one pass.
Section 05
Before AI: the readiness gate
Measures: what an AI could read today
The handful of checks that decide whether AI agents, copilots and scheduling have anything to run on right now, or whether the foundation comes first.
Prefer to have it scored for you rather than tick boxes by hand? The AI Readiness Assessment is the interactive version: twelve questions, and it returns your score, the phase you are actually in, and your next three actions on screen, free.
The full checklist
Get the full checkable list.
You have seen the five sections above. Enter your work email and the full set of concrete yes-or-no checks opens right here on this page, and a copy goes to your inbox to work through on the floor.
Every checkable item under all five sections, plain yes or no
A short note on what your answers mean and what to fix first
The one public rule that sets the clock on lot traces, with a link
Work email only. We use it to send the checklist and nothing else you did not ask for. Unsubscribe anytime.
Unlocked. The full checklist is open below, and a copy is on its way to your inbox. If you checked the box, a Harmony engineer will reach out to walk it with you.
You have checked 0 of 0 items.
Section 01 · Digitize
How much still starts on paper
Walk it station by station with a line lead. Check every one that is true on a normal production week, not a good one.
Operators fill in batch or production sheets by hand.
Quality checks are recorded on paper forms or a clipboard.
Downtime, rounds or maintenance logs are written on paper.
Shift handoffs live on paper or a whiteboard, not in a system.
The same number gets written on paper and later typed into a system.
There is at least one binder someone would pull from in an audit.
What these mean
Every box you checked here is a capture point: a place your week begins as paper, invisible to any system until a person types it in. A record written by hand does not exist to any software until it is re-keyed. These are the first thing to fix, because digitizing at the station is what the whole sequence stands on.
Section 02 · Connect
What your machines and systems share
Go asset by asset, then system by system. Most machines already know their numbers. The question is whether the numbers get out.
Machine counts come off an HMI screen or a clicker and get written down.
Faults and cycle times stay on the machine, with nothing collecting them centrally.
You do not know whether your PLCs can share data over a standard connection.
Your ERP, MES, quality and warehouse systems do not automatically share records.
A record entered in one system has to be re-entered in another by a person.
What these mean
If machine data never leaves the machine, there is no live signal to schedule against, predict failures from, or hand to an agent. If your systems do not talk, people become the integration layer and the same figure drifts apart in three places until nobody trusts any of it. Many controllers can already share data at the PLC over OPC UA, so the common case is not a plant that cannot measure. It is a plant whose measurements stop at the panel.
Section 03 · Unify
Whether your numbers are current and entered once
Two things to feel out: how old the numbers are when you act on them, and how much of the week is spent re-typing figures that already exist.
The morning meeting is about yesterday's numbers, not this shift's.
Some numbers you rely on are only complete at month end.
One person rebuilds the daily report or the schedule in a spreadsheet each day.
The plant would stall if that spreadsheet owner were out for a week.
Staff spend hours each week typing numbers that already exist somewhere else.
What these mean
Latency is the distance between something happening on the floor and anyone outside the station seeing it. It sets a hard ceiling on how proactive you can be, because you cannot correct a run you only hear about at month end. Re-keying is the most legible cost on the list, because it converts straight into a budget conversation. A spreadsheet one person maintains is a single point of failure and a source no system can read, so it can never feed a scheduler, a search tool, or an agent.
Section 04 · Prove it
Can you trace one real lot fast
Pick one real lot and trace it end to end, from question to evidence, on the clock. Do not rehearse it. The first honest attempt is the answer.
A one-lot trace means pulling binders and making phone calls.
A real trace would take hours or longer, not minutes.
The trace has to be stitched together across separate systems by hand.
You could not produce lot records within 24 hours in a sortable electronic file.
What these mean
Trace time is the most honest single test of whether your records are connected, because a trace has to cross receiving, production, quality and shipping in one pass. Anything disconnected shows up immediately as a phone call or a binder. What a trace is now expected to take is not a matter of opinion: under the FDA Food Traceability Rule, which implements section 204 of the Food Safety Modernization Act, covered firms must make required traceability records available to an authorized FDA representative within 24 hours of a request, and during an outbreak, recall, or other public health threat must provide the required information in an electronic sortable spreadsheet within that same window. A paper trace can meet neither condition reliably.
Source: 21 CFR 1.1455, paragraphs (c)(1) and (c)(3)(ii), FDA Food Traceability Rule.
Section 05 · Before AI
The readiness gate
The last few checks decide whether AI has anything to run on today, or whether the foundation comes first.
A vendor has pitched AI agents, a copilot, or forecasting into your plant.
You are not sure what data those tools would actually read.
Most of your records are not yet in a system that software can read.
You have bought AI tooling before and never found out why it did not stick.
What your answers mean
The more boxes you checked across all five sections, the closer you are to the start of Phase 1. Every checked box is a place a system, and therefore any AI, is blind today. That is not a criticism of the plant. It is a description of where the sequence begins.
None of these needs a new AI tool to fix. They need the record digitized at the station, the systems connected, and the data unified into one live layer, in that order, before AI. Failed AI purchases almost always trace back to missing or disconnected data rather than a bad model. If you want this turned into a score and a set of next actions, run the AI Readiness Assessment.
Where these checks put you: the three phases
The checklist maps onto the same three phases every plant moves through. A high count of checked boxes means you are at the start of Phase 1, whatever you have been sold. That is the most fixable position on the list, because the work is known and unglamorous rather than experimental.
Phase 1
Lay the Data Foundation · Digitization
Every pen-and-paper record digitized at the station, every software system connected, and all of the data unified into one live layer. The digital transformation starts here.
Phase 2
Production & Operations Scale
Factory operations turn proactive: live sensors and machine data, the AI scheduling board, predictive maintenance before failure.
Phase 3
AI-Native Operations
Agents across the floor and the back office act on the live layer: quality signals, reports, copilots. Humans approve.
Want the same five measures counted in person on your own floor? That is the Manufacturing Paper Audit, the long version you walk station by station. And for a dollar figure on what the gaps cost, the ROI Calculators & Tools price it on your own inputs.
Rather have someone walk the checklist with you?
The same walk is the first week of a Harmony pilot, on-site, with forward-deployed engineers doing the counting alongside your team. Phase 1 first, because that is the order it has to happen in. See what the live layer looks like.
