Service
Is your factory ready for AI?
An assessment that tells you what you already have, what is missing, and where artificial intelligence would genuinely pay off — before you commit budget to it.
Why this assessment exists
Most industrial AI projects fail before the model is written.
They fail because the dataset was never standardised, and because each department interprets the same data differently — the same stoppage is downtime to production, a repair to maintenance, and a line of free text to whoever typed it in.
The technology is rarely the hard part. The hard part is that a useful model needs consistent, granular, trustworthy data about what actually happened on the shop floor — and most factories were never built to produce that. Not through negligence: production systems get built to run production, not to feed analytics.
So the honest first question is not "which AI should we use". It is "what would we need in place for AI to tell us something we do not already know". This assessment answers that question, in your factory, with your systems.
What we examine
Seven dimensions, assessed on site.
We do not send a questionnaire. We walk the floor, sit with the people who use the systems, and look at the actual data.
Data availability
What is recorded today, at what frequency, and at what level of detail. A daily production total and a per-cycle record are different worlds.
Data quality
Whether the same event is recorded the same way twice. Free-text fields, inconsistent reason codes and retrospective corrections all limit what can be built later.
Existing systems
What the ERP, MES, quality and maintenance systems hold — and, more importantly, whether that data can be got out of them.
Integration
Where information is re-typed from one system into another. Every manual hand-off is both a delay and a place where meaning drifts.
Shop floor capture
What is measured automatically by equipment versus written on paper and entered later. This is usually the single biggest constraint.
Traceability
Whether a finished unit can be linked back to its materials, machine, operator, settings and time window.
Infrastructure
Network coverage on the floor, equipment connectivity, and what can realistically be read from older machines.
How it runs
A structured engagement, not an open-ended study.
Framing conversation
We agree what you want to be able to decide better. Without that, an assessment produces a document nobody uses.
On-site observation
We follow the production flow and see how information is created, recorded and passed on.
Systems and data review
We look at real extracts from your systems, not schema documentation. What is in the fields matters more than what the fields are called.
Interviews
Planning, production, quality, maintenance and IT. Each sees a different part of the same picture.
Analysis and scoring
We map maturity per dimension and identify what blocks each candidate use case.
Findings review
We present the conclusions and work through the priorities with you before writing the final report.
What you receive
A report you can act on, and argue with.
Maturity map
Where you stand on each of the seven dimensions, with the evidence behind each conclusion.
Candidate use cases
Concrete applications relevant to your operation, each with what it would require and what it would plausibly deliver.
Blockers
What has to be true before each use case is worth attempting — usually a data or process change, not a technology purchase.
Prioritised roadmap
A sequence, with the cheap high-value work first. Often the first steps involve no AI at all.
Risks and limits
Where AI would not help, where it would be hard to validate, and where the cost of being wrong is too high.
Possible conclusions
Three honest outcomes.
We are not paid to conclude that you need artificial intelligence. These are the three answers we actually give.
Ready in specific areas
You have the data and the process discipline for one or two well-defined applications. We show you which, and what the first one would involve.
Not yet — and here is the gap
The opportunity is real but the foundations are missing. We tell you exactly what to fix first, and that work usually pays for itself on its own.
The problem is not an AI problem
Sometimes the bottleneck is visibility, or a process that nobody owns, or two systems that do not talk. Saying so is more useful than selling you a model.
Start with the assessment, not the software.
Tell us what you would like to be able to decide better, and we will tell you whether your factory is in a position to support it.
Show us how your factory works.
Let us look at your processes, identify where value is being lost, and determine where technology can actually help.