Key takeaways
- Lights-out manufacturing is achievable – but most plants are further from it than they think.
- The barrier isn’t technology. It’s governance, digital maturity, and the foundations that sit beneath automation.
- There are nine stages of digital maturity. Not every plant needs to reach stage nine, but every plant needs to get the first three right.
- Lights-out manufacturing doesn’t mean lights-out people. Upskilling, change management, and bringing people with you is what separates projects that scale from projects that stall.
What’s the vision manufacturers are being sold?
Lights-out manufacturing. What does this term mean? Lights-out manufacturing or dark factories is a manufacturing methodology where production lines are autonomous, running with minimal or no human intervention, and enabled by full automation, robotics and AI-led decision-making.
But it’s more than a vision. Fully autonomous manufacturing is technically possible and, for certain applications, already happening. However, there’s a gap between the vision and the reality that shouldn’t be underestimated.
So, the question isn’t whether lights-out manufacturing is achievable. It’s what it really takes to get there – and whether your plant has the foundations in place to make the journey.
Where are most manufacturers in reality?
At ITI Group Consulting, we work with manufacturers at every stage of their digital maturity. There are nine stages of this model that map the journey from basic data collection through to full operational autonomy.
The first three stages are foundational. Stage one is where many organisations start: you have a lot of data, but it’s not accessible. It’s locked in systems that don’t connect, stored in formats that aren’t usable, or simply not being collected consistently. Stage two is about getting that data stored centrally. Stage three is about turning it into real-time, visible insights.
Most manufacturers we work with sit somewhere in these first three stages. Some are further along in specific areas – for example, a pilot that’s working well in isolation – but as an organisation, the foundations aren’t yet consistent or connected.
More importantly, these early stages aren’t something to rush through or treat as a tick-box exercise. Getting these foundations right is crucial to continuing on your digital maturity journey. After all, you can’t automate chaos.
Not every plant needs to reach stage nine. That depends on your business, your operations, and what you’re trying to achieve. Two manufacturers in different sectors will have very different requirements, but every plant needs to know where it really sits on the spectrum before deciding where it’s going.
Why does governance matter more than technology?
When organisations set a lights-out manufacturing goal, they often focus on the technology required – for example, which MES platform, or which AI tool – but the thing that actually determines success or failure is far less glamorous: governance.
Take a global food and beverage manufacturer that set out to modernise its manufacturing execution capability across multiple European production sites. They wanted to standardise execution, improve traceability, and reduce manual dependency. On paper everything was set up for success; in reality, the programme stalled over the next two years.
This was because the organisational foundation was not in place. So, different sites interpreted the programme in different ways, leading to fragmentation as some sites built local tools for local problems, while others waited for centralised direction.
This direction never came because there was no clear decision-making authority. Consequently, questions about system ownership, configuration standards, and integration boundaries were lost in a messy pipeline, or were not resolved at all.
Lack of centralisation also meant that every site became its own design authority resulting in no standardisation, and growing divergence from the original intent. Meanwhile, the same data was captured by multiple systems and integration requirements shifted as boundary decisions changed beneath the architecture.
The cumulative effect was that confidence in the programme began to erode, as delivery timelines slipped and OEE improvement targets were deferred. What was needed was a design authority, a centralised framework with scope for local differentiation, and sequenced decision-making.
The technology was never the constraint – governance was.
We saw a similar pattern with a manufacturer undertaking a greenfield MES implementation as part of a wider digital transformation programme. There were multiple system integrators working on site in parallel – ERP, MES, automation and OEM vendors. However, there was no formal mechanism for resolving cross-vendor design decisions.
Therefore, the project appeared to be progressing, but it wasn’t. When asked to produce a functional design specification and cost estimate, the team had 38 open design decisions blocking delivery. Any figure produced carried an accuracy of ±40-50%, which was too wide to be commercially meaningful.
The lesson is consistent across both programmes: client governance is not an internal concern – it is a delivery dependency. Design decisions need a named owner, a resolution deadline, and a clear escalation path. The structure that resolves design decisions in week two is the same structure that prevents programme failure in month twelve.
What about the people?
Lights-out manufacturing shouldn’t mean lights-out people.
The excitement around autonomous production sometimes creates an assumption that the goal is to remove people from the equation entirely. It isn’t. Technology is exciting, and automation is powerful, but you still need people for change management, for problem-solving, and for the judgment calls that algorithms can’t make.
Lights-out manufacturing marks a move towards people doing different, higher-value work, which requires investment in upskilling: digital and data fluency, systems and automation literacy, and an analytical and continuous improvement mindset.
And crucially, you need to take people on the journey with you, so that they understand and trust in new systems and ways of working, and they comprehend the impetus behind the change and the desired outcome. We’ve seen a pharma manufacturer invest heavily in an OEE system, only to find their operators still calculating on spreadsheets. We’ve seen scheduling tools powered by sophisticated algorithms, where the schedulers ultimately ignore the output and drag and drop their own plan anyway.
The technology works. But if your people aren’t with you – if they see the transformation as a threat rather than an opportunity – your project will stall. Not because of the technology, but because the culture isn’t ready for it.
Where should you start?
The journey to lights-out manufacturing is a sequence, and each step matters.
1. Strategy definition
Start by clarifying what lights–out manufacturing really means for your organisation. For example, what are the top three to five pain points costing you time and money? Quantify the impact of each of these, prioritise them based on feasibility and business value, and then build a change management plan before you touch any technology.
2. Current state assessment
You can’t automate what you can’t see. Get an honest picture of where your plant sits on the digital maturity spectrum – not where you think it sits, but where it is in reality.
3. Connectivity and infrastructure review
Digitalisation starts with a strong infrastructure foundation. This means reviewing your network architecture, your OT and IT segregation, your connectivity across the shop floor. This is the ground everything else is built on.
4. Pilot initiation
Demonstrate measurable value and build internal momentum. Choose your pilot carefully – look for initiatives that score well on value (will it make a tangible difference?), visibility (will people across the organisation see the result?), and viability (can you actually deliver it with the resources you have?).
5. Scale
Scaling isn’t just doing the pilot again in more places. It requires a deliberate decision about what gets standardised across sites and what stays locally owned. Standardisation reduces your total cost of ownership by simplifying your architecture across sites. It reduces business continuity risk by embedding knowledge into workflows rather than leaving it in people’s heads, and it gives you real–time visibility across your whole network. The manufacturers who scale successfully treat their pilot as a template – not just a proof of concept.
6. Autonomy
Once everything beneath it is solid – the data foundations, the governance, the connected systems, the skilled workforce – then autonomy becomes achievable. AI and machine learning for predictive and prescriptive optimisation; digital twins for simulation and process tuning; closed-loop control where algorithms adjust parameters directly. This is lights–out manufacturing – and it’s built on every step that came before it.
Is your plant ready for what comes next?
The path to lights-out manufacturing is real, but it’s longer than most vendors will tell you – and it starts well before you buy anything. Get the strategy clear, get an honest read of where you are today, and get the governance in place before the first line of code is written. The technology will follow. The foundations won’t build themselves.