DARK FACTORY CONSULTING — LIGHTS-OUT TRANSFORMATION STRATEGY

Once AI starts delivering results,the next place is a factory that runs itself.

Long-term strategy consulting that transforms a whole plant into a lights-out dark factory run by AI. We design it as a staged journey — foundation, integration, intelligence, autonomy — with autonomous-control governance that hands over control only as far as it is verified, and a wave structure that invests stage by stage. The starting point is the results earned in Manufacturing AX Consulting.

FIELD 200+ SITES / 16 YRS · AI AUTONOMOUS FACTORY ×2 · AI→PLC DIRECT CONTROL VERIFIED · L3 → L4

01PROBLEMFROM AI PILOTS TO AUTONOMOUS OPERATION

You've delivered results —
but whole-plant lights-out is hard to even begin.

Adopting individual AI and running a whole plant autonomously are different problems. On the road to lights-out operation, the wall companies hit is not data — it is sequence, safety and investment.

You don't know how far lights-out can go

Every process has different conditions for autonomy to hold. There's no basis for judging what can go unmanned, and by when.

Skip the order and it runs aground

AI before standards, control handover before verification — autonomy projects that skip the order are the ones that most often fail.

The safety boundary of autonomous control is daunting

When you hand equipment control to AI, adoption itself is dangerous without a safeguard that physically blocks a wrong command.

You can't invest it all at once

A whole-plant transformation is multi-year and large-scale. A structure that ties up big money upfront on an unverified stage is hard to accept.

02WHAT IT ISMASTERPLAN · CONTROL GOVERNANCE · WAVE INVESTMENT

Here's what Dark Factory Consulting does.

WACE's manufacturing-AI engineers diagnose your plant's current maturity and design the goal of whole-plant lights-out operation and the stage-by-stage transformation roadmap to get there. The entry stage — data standardization and turning AI into results — is handled by Manufacturing AX Consulting; on top of that, this line owns the transformation strategy toward L3→L4 autonomous operation — with final delivery continuing into Autonomous Factory Build.

A factory transforming stage by stage into autonomous operation

FIG — L3 → L4 · STAGED TRANSFORMATION

Transformation masterplan

We diagnose current maturity to define the lights-out goal and design the stage-by-stage roadmap — foundation → integration → intelligence → autonomy. We set the basis for judging which process can go unmanned, and by when.

ROADMAPDIAGNOSIS → GOAL DEFINITION → STAGED ROADMAP

Autonomous-control governance

We hand control to AI, but a permission framework (L0–L4) and physical-constraint verification physically block a wrong command. Control is handed over only as far as verified, within a safety boundary.

SAFETYAAF PERMISSIONS L0–L4 · SHACL PHYSICAL-CONSTRAINT BLOCKING

Wave-separated investment structure

We split the whole-plant transformation into stages (waves) for ordering and investment. You confirm each stage's result before deciding the next investment, so no big money is tied up upfront on an unverified stage.

INVESTSTAGE-SEPARATED ORDERS · GATE PASSED = INVESTMENT RELEASED

VERTICAL

The final deliverable isn't one plant's lights-out — it's a repeatable system of transformation

We design the data standards, the ladder of verification and the judgment the system has learned so they can be transplanted to the next line and the next plant. So the first plant's transformation stays as an asset you can reuse again and again.

THREADDATA STANDARDS → VERIFICATION LADDER → LEARNED JUDGMENT — TRANSPLANTED TO NEXT LINE · PLANT

03IN ACTIONREAL PRODUCT SCREENS

After adoption, this is how
the factory works.

Once data governance is in place, this is how the way the factory works changes. Below are real VEXPLOR screens.

D-01 · Manufacturing

Automated predictive maintenance

AI catches early signs of equipment trouble in real time and alerts the right person the moment risk crosses a threshold.

STACK — PREDICTIVE MAINTENANCE LSTM · ANOMALY DETECTION · THRESHOLD ALERTS

D-02 · Manufacturing

Automated root-cause analysis for quality

AI cross-checks process data to flag defect patterns early and presents cause and response with the evidence attached.

STACK — 4-WAY RAG · ONTOLOGY EVIDENCE CHECK

D-03 · Manufacturing

AI → direct equipment control

It syncs digital-twin predictions with the live floor and pushes AI decisions down to the PLC to control equipment directly.

STACK — 3D TWIN · AI→PLC DIRECT CONTROL VERIFIED

D-04 · Manufacturing

Automated floor and production reporting

AI automatically compiles daily and weekly output, utilization and inventory, adds a summary, and sends it at a set time.

STACK — SHOP-FLOOR POP · AUTO-COMPILED REPORTS

Screen data is shown for illustration — in the autonomous stage, people intervene only on exceptions, and unmanned run time and intervention counts are measured every month

04THE JOURNEY4 STAGES · GATED · MULTI-YEAR

The road to the dark factory —
four big steps.

This is not overnight unmanned operation. It is a staged journey — lay the foundation, connect it into one, build up the intelligence, then hand over control — with a gate between every stage, its passing criteria agreed in advance. Skipping this order — adopting AI before data standards — is the most common reason manufacturing-AI projects run aground.

FOUNDATION

A standard for equipment and data

We set one standard all plant data will follow, and write it into equipment purchase contracts. This is where the rework risk of every later investment is removed.

GATESTANDARD RATIFIED · DATA QUALITY CRITERIA MET — JOINT VERDICT

INTEGRATION

One connected, visible factory

From floor equipment to the boardroom, data flows without passing through human hands — and the numbers in the business systems finally agree with the numbers on the floor.

GATECOLLECTION COVERAGE · TRACEABILITY — MEASURED

INTELLIGENCE

A factory that explains and predicts

Root-cause analysis, prediction, virtual verification — the system takes over the veteran's judgment, evidence attached. Breakdown repair becomes planned maintenance.

GATEEXPLANATIONS PASS FIELD VERIFICATION — NO CONTROL HANDOVER BEFORE THIS

AUTONOMY

Handing over control, step by step

From shadow verification to unmanned hours, control is handed over only as far as verification has reached. Unmanned run time and intervention counts are reported monthly, in numbers.

GATEQUANTITATIVE CRITERIA AT EVERY STEP · JOINT VERDICT

FOUNDATION → GATE → INTEGRATION → GATE → INTELLIGENCE → GATE → AUTONOMYFIG — STAGED JOURNEY · TYPICALLY A JOURNEY OF SEVERAL YEARS, NO FIXED TERM PROMISED — YOUR STAGES, GATES AND TIMELINE ARE SET IN THE DIAGNOSIS ROADMAP

DESTINATION

Lights-off hours become the default, and people are called only for exceptions.

Through nights and weekends, the equipment keeps itself running; people's work moves from performing tasks to monitoring, handling exceptions and finding the next improvement. And the final deliverable is not one plant running unattended — it is a system of transformation your company can use again and again: the data standards, the ladder of verification and the judgment the system has learned all carry over to the next line and the next plant. This destination is not a promise; it is reached only through the gates, passed one by one.

END STATEUNMANNED RUN TIME · INTERVENTION COUNT — MEASURED MONTHLY · A REPEATABLE SYSTEM OF TRANSFORMATION

No gate, no next stage

Criteria are agreed as quantitative metrics in advance, and the verdict is never our own declaration — it is confirmed jointly by your management and our lead.

You never buy it all at once

Contracts and orders come one stage at a time, as the roadmap defines them. You see each stage's result with your own eyes before deciding the next investment.

05DIAGNOSIS8-WEEK PROGRAM · WEEKLY REVIEW · ENGINEERS ON-SITE

The 8-week diagnosis — the final report's
contents are agreed in week one.

Every journey starts with a diagnosis. The standard diagnosis-and-design program runs eight weeks, carried out on-site by WACE manufacturing-AI engineers. The timeline adjusts to the state of your plant (and is fixed in the diagnosis), but the skeleton — and the moments where you check in — stay the same.

Kickoff

We align on goals and success criteria, and agree the table of contents of the final report together. From week one, you know what will come out at the end.

EVERY WEEK

A 45-minute weekly review

A fixed day, a fixed format — progress, new findings, issues that need a decision, next week's plan. There is not a single week when you don't know where the project stands.

Interim report

Your current position and the list of broken links blocking autonomy — confirmed together with management.

Individual pre-briefings

We brief each key executive in person beforehand — so nothing in the final report is heard for the first time.

Final report & handover

The final report, together with the execution roadmap and the operating handover.

W1 KICKOFF → WEEKLY 45-MIN REVIEW → W4 INTERIM REPORT → W7 PRE-BRIEFINGS → W8 FINAL REPORT · HANDOVERWHAT YOU RECEIVE — DIAGNOSIS REPORT · INTEGRATION DESIGN · EXECUTION ROADMAP · 1-PAGE EXECUTIVE SUMMARY · OPERATING HANDOVER PACKAGE / STAFFING: 1–3 WACE MANUFACTURING-AI ENGINEERS ON-SITE

06HOW WE WORKQUALITY PROMISE ×4 · PAY AFTER PROOF

How we work — you pay after
you've seen the result.

We work by predictable rules. Four quality promises we keep, and the roles we ask of you to make the project succeed.

Every number has a source

Any figure in a report can be traced in a footnote to which data it came from, and how.

Diagnostic figures are reproducible

For figures computed from your data — utilization, defect rates — we hand over the calculation as a script. Only when you can reproduce the same value from the same data does a diagnosis become measurement rather than assertion.

No surprises

The substance of interim and final reports is shared individually before the meeting. The worse the news, the earlier it arrives — and in writing.

Three honest options when a gate is missed

Remediate and re-judge, proceed with only what passed, or stop and settle — agreed with you and put on paper. We do not paper over a miss and move on.

GATE PASSED = STAGE ACCEPTED = INVESTMENT RELEASED — YOU PAY AFTER YOU'VE SEEN THE RESULTSTAGE COMPLETION IS JUDGED JOINTLY AGAINST PRE-AGREED CRITERIA · INVOICING IS TIED TO GATE PASSAGE

YOUR TEAM

Executive sponsor

One executive to join the monthly steering meeting and approve the stage gates.

Project owner

One working-level lead to attend the weekly reviews and coordinate internally.

Data authority

Someone to approve the provision and release of system data.

Site liaison

Someone to arrange interviews and floor inspections.

Recommended — one or two of your engineers join the project and do the analysis with us. Not observing but sharing the actual work: the surest way the capability stays with you

07STANDARDS12 PLAYBOOKS · 339 CHECKS · CERTIFIED BEFORE STAFFING

The same quality, whoever shows up —
we work from documented standards.

Consulting that rests on individual brilliance changes quality when the consultant changes. We run every activity of the diagnosis from documented standards — whoever is staffed, the same order, the same criteria, the same quality.

12 activity playbooks

From pre-diagnosis to closeout and handover, every activity has its procedure written down — down to the standard lines a session opens with.

SCOPEPRE-DIAGNOSIS → KICKOFF → INTERVIEWS → VERDICTS → WEEKLY REVIEW → REPORTING → CLOSEOUT · HANDOVER

339 execution checks

Every activity is run against a checklist — and the check record itself is the evidence for stage acceptance.

EVIDENCECHECK RECORD = GATE EVIDENCE

Certified before staffing

Every consultant completes training and a pass-required assessment before joining a project, then supervised field work before leading alone.

GATEPASS-REQUIRED ASSESSMENT · SUPERVISED FIELD TRAINING

A reproduction script with every figure

Every diagnostic figure ships with the script that computed it. You can re-measure the same value at any time.

REPROSAME DATA → SAME VALUE — ANY TIME

The whole system is updated after every project — your project's experience becomes the next standard

08FIT & ESTIMATE4 PROFILES · 3-AXIS · NO LIST PRICE

It fits companies like these —
and here is how we price it.

Ready for autonomous operation

Manufacturers already delivering results from a smart factory and AI adoption, preparing for the next stage — autonomous operation.

Lights-out masterplan

Companies that need a whole-plant lights-out roadmap and an autonomous-control governance system.

Wave investment

Companies that want to pursue the dark-factory transformation carefully, with a stage-separated investment structure.

First entry

If you're still at the diagnosis-and-results stage, start with Manufacturing AX Consulting.

SCOPE

What's included

Data and floor diagnosis · AI use-case design · build and integration · closeout report.

TEAM1–3 WACE MANUFACTURING-AI ENGINEERS · TIMELINE SCOPED TO YOU

FORMULA

Three pricing axes

Data scale and integration breadth / scope of use-case design (number of AI use cases) / build and integration depth.

3-AXISDATA SCALE · USE CASES · DEPTH

PACKAGE

Bundled discount

We scope an exact quote after a meeting, and bundling consulting with training earns a discount.

OPTIONGROUP-TRAINING BUNDLE DISCOUNT · 1:1 FACILITATION EXTRA DISCOUNT — GROWS AS YOU BUNDLE

Pricing page — see the full basis →

09WHY WACEDOMAIN DEPTH × STANDARDS × PRODUCT

There's a reason it has to be WACE.

Manufacturing depth, proven

16 years and 200+ manufacturing-site engagements. AI→PLC direct control verified, two government-backed AI autonomous factory builds underway — territory a generalist AX vendor struggles to enter.

PROOF200+ SITES / 16 YRS · AI→PLC VERIFIED · AUTONOMOUS FACTORY ×2 SITES

Built on national standards and patents

We lead the consortium developing Korea's manufacturing-data standardization reference model (AAS/IEC 63278), hold 2 registered and 1 pending manufacturing-AI patents, and run a corporate R&D lab.

IPPAT REG 2 + PENDING 1 · IEC 63278 · R&D LAB

Product-backed consulting

We consult on top of a Manufacturing AI OS with 20 industrial solutions, 834+ standard tables and a manufacturing-tuned 4-Way RAG, so diagnosis flows straight into build and operation.

BASESOLUTIONS 20 · TABLES 834+ · 4-WAY RAG

CASE-15%Defect rate — Company H (die casting)

CASE99%Inventory accuracy — Company K (automotive electronics)

EXPECTED-90%Data processing time

EXPECTED-15%Process operating cost

And this is consulting whose methodology exists as documents — the staged journey (04), the 8-week diagnosis (05), the quality promise (06) and the 12 playbooks (07): how we work is open before you sign

Resources — download the Dark Factory Consulting brochure →