AIXEL Northwind Components - Digital maturity model & Diagnostic

Digital maturity model & Diagnostic

Prepared by AIXEL for Northwind Components · 2026-07-25 · SMB industrial manufacturer (components) (EU) · Acatech Industrie 4.0 Maturity Index . Gartner Analytic Ascendancy.
HOW TO READ THIS DOCUMENT

Digital transformation is not a single feature: it is a path that advances in stages.

This document does two things
1A reference model
how the information system of a components manufacturer should be built and how it evolves.
2A diagnostic
the questions that let us, together, see which rung you are on today.
How we use it
The modelYour reality (the questions)The gap = your roadmap
guides the choices:buildbuyintegrate
The through-line: The ambitious capabilities - predictive maintenance, connected machines, real-time margin - live on the upper rungs. Those rungs do not stand without a unified core and a live shop-floor data layer beneath them. Here is the whole ladder, here are the questions: let us find which rung you are really on and build the path up.

Variable #1 (the unknown that dominates everything): the shop-floor / ERP data spine (C2) - whether machine, order and quality data resolve into one record, or live in per-line silos reconciled by hand. It gates everything above and is the priority question.

The model rests on recognized, validated standards (not invented). This is an automated preliminary screening from public information only - the observations are hypotheses to confirm together.

1
Our role: architects, not feature vendors
A feature vendor sells you...AIXEL - architect
"here is an MES / here is a dashboard""here is how your whole operating and data system should look, and how to get there in sequenced waves"
sells a productsells a destination map and the guidance along it
starts from a featurestarts from the maturity level and the honest gap
Your hidden asset

You already run real production with real data being generated on every line. Today that data is not unified. When it is, it becomes the defensible asset that lets you scale margin, not just volume.

Market context. The sector leaders now compete on data and connected operations, not headcount. A unified data layer is what lets a manufacturer grow while holding margin.

2
The reference: target architecture

Six layers bottom-up. The upper layers do not function without the lower ones: that is the whole logic of the sequence.

1
System of Record . Operating core (C1)
ERP + MES: orders, production, quality, inventory. The foundation. Adopt/configure, do not build.
2
System of Integration . Data layer (C2)
one record across machines, orders and quality; the single source of truth on which scaling depends.
3
System of Engagement . Customer/supplier front door (C3)
portals, structured quote/order flow, self-service. Where demand meets the operation.
4
System of Relationship . CRM (C4)
lead-to-account lifecycle, retention, cross-sell across the customer base.
5
System of Intelligence . Analytics (C5)
descriptive to predictive: OEE, margin per line, forecasting on the data layer.
Spine A . AI & automation (C6)
predictive maintenance, quality vision, planning automation - at the seams, human-in-the-loop, model-risk governed.
Spine B . Governance & trust (C7)
security, GDPR, OT/IT segregation, data residency. A requirement at every layer, not a stage.
Reference (best-in-class, to ground how it should be)

The sector's connected-operations leaders run a unified data layer feeding predictive maintenance and real-time margin. For an SMB these are aspirational: they define the ceiling (L4). The realistic near-term destination is Connected (L2).

3
The maturity ladder L0-L4

The five rungs are anchored to recognized frameworks (the standard column), validated on the primary source.

LvlNameWhat it is≈ Acatech I4.0
L0FragmentedSiloed tools / manual; no unified data; paper or spreadsheets on the floorComputerisation
L1StandardizedUnified ERP/MES core; electronic documentation; digital basics across linesConnectivity
L2ConnectedLive data layer; one record across machines/orders/quality; a 360 viewVisibility
L3IntelligentAnalytics/AI embedded in decisions; predictive maintenance; managed BITransparency / Predictive
L4Predictive / PlatformPredictive support in production with model-risk controls; each new line plugs inAdaptability
The key point of the conversation

The ambitious capabilities sit at L2-L4, but they rest structurally on L1 (a unified core + structured data). The honest first question is not which feature you want, but which rung you are really on.

4
Where you are now: maturity map

current state (hypothesis)   target   to be defined

DimensionL0L1L2L3L4ConfidenceDist.
C1 · Operating core (ERP/MES)Medium1
C2 · Data & interoperability (single source of truth) ★to be defined~1.5
C3 · Access & engagement (customer front door)High1
C6 · AI & automation (spine)Low~1.5
O1 · Digital strategy & leadershipMedium1

Weighted composite: L0.8L2.0 · weighted total gap ~1.2 rungs, concentrated on the gating foundations (C1, C2). Reported as a hypothesis, not a grade.

Overall reading

A real operation generating real data, with a genuine strength in the customer front door (C3), sitting on foundations of unknown, probably-low integration (C1/C2). The honest destination is Connected (L2): unify the data and prove margin - not AI-native. The gap is modest and foundation-shaped.

5 · The dimensions: levels, questions, reading

C1
Operating core (ERP/MES)
weight 3 · L1
L0
Spreadsheets + disconnected tools per line
L1
A standard ERP/MES exists but per-site
L2
One core across all lines, one record
L3
The core feeds analytics and is automated
L4
Platform: each new line plugs into the core

Questions

  • What single system, if any, does a job flow through from order to delivery to invoice?
  • Are the lines/sites on one ERP/MES or separate systems?
  • Can you show the tool where a live job is managed, and export one job end-to-end?
Preliminary reading . to confirm

Probably L1. Digital operating systems exist, so above L0 - but no evidence of a single core spanning all lines.

Target: L2 (one core is required by the scaling thesis; not L4) · Gap: 1 · Confidence: Medium · Confirm if: at the call, seeing one system a job flows through from order to delivery to invoice.
C2
Data & interoperability (single source of truth)
weight 3 · to be defined
L0
Data siloed per machine/tool; reconciliation manual
L1
Some shared references; systems do not talk
L2
Integrated: one record across the operation
L3
The layer feeds real-time decisions
L4
Platform data layer; new lines auto-onboard

Questions ★

  • When you run an order, do machine, order and quality data live under one record or in separate tools?
  • How is a job's data reconciled across the shop floor, the ERP and quality?
  • Is there a data model / systems diagram? Can you export one job end-to-end?
Preliminary reading . to confirm

One of the two decisive unknowns. No integration platform observable; machine, order and quality data likely live in silos. Gating + low confidence: no level asserted.

Target: L2 (single source of truth is the precondition for everything above and for scaling) · Gap: ~1.5 · Confidence: to be defined · Confirm if: whether a job's machine data, order and quality result can be seen against one record.
C3
Access & engagement (customer front door)
weight 1 · L2
L0
Phone/email only; brochure site
L1
Contact forms + content
L2
Online quote/order flow + account portal
L3
Omnichannel + messaging + a real app
L4
Proactive, orchestrated journeys

Questions

  • How many orders actually come through the online flow vs by phone/email?
  • Is the customer portal used, and for what?
  • Can you share online-order volume and portal active-user stats?
Preliminary reading . to confirm

Probably L2 (High). A working quote/order flow and an account portal were observed - a stronger front door than most SMB manufacturers have. Credit it.

Target: L3 (add two-way messaging + a customer app; justified by the existing L2 base) · Gap: 1 · Confidence: High · Confirm if: whether the portal is actually transacted-through or a shell.
C6
AI & automation (spine)
weight 1 · L0-L1
L0
No AI/automation; all manual
L1
Awareness: ad-hoc tool use, no governed workflow
L2
AI/automation embedded at one seam (governed)
L3
AI across processes with model-risk controls
L4
AI-native operations

Questions

  • Where is AI used today, and who checks the output?
  • Which processes are the most manual and repetitive (automation candidates)?
  • Are there internal policies on AI and data?
Preliminary reading . to confirm

Probably L0-L1 (Low). Staff plausibly use gen-AI tools ad hoc; no automation platform or governed AI workflow observable.

Target: L2 (predictive-maintenance and planning automation are high-value, low-risk once data exists; L3/L4 not justified) · Gap: ~1.5 · Confidence: Low · model-risk: before L3/L4: a validation protocol, drift monitoring, and human-in-the-loop sign-off for any predictive/quality model · Confirm if: whether any process has AI embedded with a defined owner and a review step.
O1
Digital strategy & leadership
weight 2 · L1
L0
No strategy; reactive
L1
Digital = a project, founder-led; no data owner
L2
A named digital/data owner + a funded roadmap
L3
Digital in the operating plan; measured
L4
Digital/AI is the operating model

Questions

  • Who owns your own data/systems (not customers'), and is there a plan to improve them?
  • Is there a digital roadmap with a budget line?
  • How are internal digital initiatives funded?
Preliminary reading . to confirm

Probably L1 (Medium). No visible owner or roadmap for the company's own data/digital maturity; digital shows up as isolated projects, not an operating capability.

Target: L2 (name a data/digital owner + a light internal roadmap; modest, high-leverage) · Gap: 1 · Confidence: Medium · Confirm if: whether anyone owns 'our own data/systems maturity' with a plan and budget.
6
From diagnostic to roadmap: waves, economics, risks

The distance from the model is your roadmap. Priority follows the dependencies: first what unblocks the rest (core + data layer), quick wins in parallel.

Wave 0

See behind the floor

  • Resolve the C1/C2 unknowns: systems and data-record inventory
  • Deliverable = the true current state
Wave 1

The data layer

  • Converge on one ERP/MES core buy
  • Stand up one record across machines/orders/quality build
Wave 2

Connect & measure

  • OEE + margin-per-line dashboards integrate
  • CRM on the customer base buy
Wave 3

Intelligence

  • Predictive maintenance (governed) build
  • Planning automation at the seams

Build / buy / integrate

StepChoiceRationale
ERP/MES core, CRM, BI, dashboardsbuycommodity best-of-breed; do not build
The unifying data layer + shop-floor integration gluebuildthe differentiating connective tissue - the architecture, not a tool
AI/predictive workflowsintegratecompose off-the-shelf models into governed workflows

Anti-patterns to avoid: (1) buying one MES and calling it the platform - the value is the integration; (2) building the core from scratch - buy a proven one.

The value case (illustrative, no prices)

Cost-of-current-state levers, to model on your data: unplanned downtime = hours x loaded line rate; scrap/rework = rate x volume x cost; manual reconciliation = hours/week across tools. Wave 1 (the data layer) is the enabler that all later waves stand on. (illustrative, to model on your data - no prices)

Key risks

Key risk: the data foundation (C2) is unknown - the whole roadmap is provisional on it, so Wave 1 starts with an inventory before any build. Any predictive/quality AI needs a validation basis before it touches production.

7 · Interventions you can start now

In parallel with the roadmap

In parallel with the roadmap, three concrete interventions that touch little or none of the deep foundations, are independent of Variable #1, and each seed a real wave.

AOEE + margin-per-line dashboard
What it is
A directional dashboard on representative then real data.
What it solves
The lack of a group-level view of load and margin.
What you will see
Your operation, finally readable.
How & when
Low dependence on the data unknowns; seeds Wave 2.
BDowntime + scrap capture
What it is
A lightweight capture of downtime and scrap at the line.
What it solves
The biggest silent margin leak, currently unmeasured.
What you will see
Where the money actually leaks, quantified.
How & when
Independent of the core; fast; seeds Wave 1 data.
CAI-assisted quote/spec drafting
What it is
A governed gen-AI workflow for quotes and specs.
What it solves
Slow manual quoting throughput.
What you will see
Faster quotes, with human sign-off.
How & when
Composable with existing tools; low risk; seeds Wave 3.
CriterionA . DashboardB . CaptureC . AI quoting
Independent of the data unknownshighhighhigh
Visible valuehigh (margin view)high (leak found)high (throughput)
Roadmap waveWave 2Wave 1Wave 3
The next step

The fastest way to make this concrete is a working session on one real line: we show where margin leaks and map the first wins together.

Book a working session
8
Confidence, limits, methodology & sources
How we assessed this

This is an automated preliminary screening from public information only (the website + independent web research). Every current-state observation is a hypothesis with a confidence tag, and each claim traces to a source. The reference model rests on recognized, validated standards.

Limits. We could not see behind the shop floor: the operating core (C1) and the data layer (C2, the Variable #1) rest on inference and absence of public evidence - so they are left to be defined and are the priority questions. The assessment is a point-in-time snapshot.

Additional references and the market-vendor map available on request, to support the build/buy/integrate choices.

9
Glossary

Core & data

ERP
Enterprise Resource Planning - the system that runs orders, inventory and finance.
MES
Manufacturing Execution System - the shop-floor production system.
OEE
Overall Equipment Effectiveness - the standard measure of line productivity.
OT / IT
Operational Technology (machines) vs Information Technology - kept segregated for security.

Method

L0-L4
The five maturity rungs, from Fragmented (L0) to Predictive/Platform (L4).
Variable #1
The single unknown that dominates the architecture - here, the shop-floor data spine.
Prepared by AIXEL for Northwind Components - 2026-07-25 - Automated preliminary screening from public data - to confirm together (not a commercial offer; no prices).