The problem with training as a separate event
Remanufacturing an engine is skilled, sequential work. Strip the block, inspect each component, rebuild to tolerance — and do it reliably, at volume, on a commercial schedule. For a site like Autocraft Engines in Grantham, one of Europe's largest independent engine remanufacturers, keeping that knowledge current across a workforce is a constant operational pressure.
The conventional answer has always been to pull people out. Send them to a classroom, run a session, return them to the line. For a large employer with deep training budgets, that is manageable. For a manufacturing SME running tight margins and leaner staffing, releasing experienced workers for formal instruction creates a real cost: the line slows, output drops, and the people best placed to train others are the ones most needed on the floor.
Lincolnshire's manufacturing sector compounds this. Skills shortages and an ageing workforce mean there are fewer workers to spare in the first place, and less margin for the kind of structured rotation that traditional upskilling assumes.
The question Autocraft's approach begins to answer is a practical one: what if the training did not require anyone to leave the workpiece at all?
What AR headsets and digital twins actually do
Picture a worker standing at an engine block, headset on. Through the visor, the physical engine is still fully visible — hands, tools, metal and all. But overlaid onto it, in precise registration with the actual components, are step-by-step assembly instructions: the correct torque sequence for a cylinder head, the clearance tolerance on a bearing, a highlight showing which bolt comes next. That is augmented reality in practice. Unlike virtual reality, it does not replace what the worker sees; it adds to it. The physical task and the digital guidance occupy the same field of view at the same moment.
The digital twin is what makes this possible in context. Rather than a generic manual or a flat diagram, the headset draws on a computational model of the specific engine assembly — a virtual counterpart that mirrors the real build and can be stepped through stage by stage, matching the worker's actual progress. A full digital twin maintains a live, bidirectional connection to the physical object, so the model and the component stay synchronised throughout the rebuild.
Combined, the two technologies project instruction directly onto the workpiece, contextualised to the exact stage of assembly in front of the worker at that moment — not a reference document to consult, but a layer of guidance built into the task itself.
The underlying idea is not new. Boeing was using early industrial AR to support aircraft wiring assembly in the early 1990s. What remains limited is small-manufacturer deployment: most of the field's 30-year history has played out in large aerospace and automotive plants. A remanufacturer of Autocraft's scale adopting these tools in a Grantham factory is, by that measure, still notable.
How Autocraft put this to work in Grantham
Autocraft Engines sits on the edge of Grantham — a company that remanufactures petrol, diesel, and hybrid powertrains for OEMs and the aftermarket, and by volume ranks among Europe's largest independent operators in that field. Its output is industrial-scale, its work precision-dependent, and its workforce must stay current as engine technology evolves without pausing commercial throughput to do so.
The company's route into AR-guided training ran through the UK government's Made Smarter adoption programme, which funds manufacturing SMEs to trial digital technologies including AR, IoT, and digital twins. That funding context matters: it is one of the mechanisms through which a mid-sized Lincolnshire remanufacturer could absorb tooling that has historically been confined to large aerospace and automotive plants.
Through the programme, Autocraft deployed AR headsets to guide workers through engine assembly procedures directly at the workstation. Step-by-step overlays — drawn from digital twin models of the specific assemblies being rebuilt — appeared registered onto the physical engine in front of the worker. New engine variants or revised procedures could be introduced to the line without pulling experienced workers off it for separate instruction. The knowledge arrived at the workpiece, in context, as part of the job rather than in advance of it.
Why learning at the workpiece changes the economics
The separation of training time and production time is one of manufacturing's oldest structural assumptions — and one of its most expensive. Every hour a worker spends in a classroom or simulation bay is an hour not spent on the floor; every scheduled retraining block carries a cost that HR budgets absorb but production schedules feel directly.
Industry-wide evidence on AR-led manufacturing training suggests this boundary can be substantially compressed. Across automotive and aerospace studies, reductions in training time of 30 to 50 per cent are regularly cited compared with classroom or paper-based methods, alongside measurable falls in assembly error rates. These are not Autocraft-specific figures; they reflect the broader field and should be read as indicative rather than confirmed for the Grantham site. But they point to a consistent pattern: instruction that arrives in context, at the moment of physical need, tends to be retained and applied more reliably than instruction delivered in advance of the task.
The reason is structural, not merely logistical. When a worker consults a manual or attends a training session, they must bridge a gap between the abstract and the actual — translating what they learned, in a different environment, into what they are now holding in their hands. AR collapses that gap. Guidance appears on the physical component at the precise stage where it is needed. Knowledge transfer becomes iterative and ambient rather than episodic: something encountered at the workpiece, repeatedly, in context — rather than front-loaded and expected to carry.
That shifts not just the economics of training, but the underlying structure of how knowledge moves through a workforce.
What this means for a place like Grantham
Geography shapes options. Grantham and the wider South Kesteven district sit at a practical remove from the engineering faculties and further education colleges that concentrate around Leicester, Nottingham, and Sheffield. That distance is not disabling — but it does mean that off-site training provision is harder to access regularly and more disruptive to arrange than it might be for a manufacturer closer to a city campus.
The East Midlands manufacturing workforce also skews older than the national average, and that demographic carries an implication beyond retirement planning. Experienced workers on an engine line hold a form of knowledge that does not transfer easily through formal curricula: the feel of a component, the judgment call made at step four that only becomes legible after doing it fifty times. Conventional training rarely captures this tacit layer reliably, and it can leave with the person who held it. AR-guided instruction offers a partial answer — not because the headset can replace that judgment, but because it can encode and distribute the procedural knowledge that underpins it, making it available to the next person on the line without requiring a structured course to deliver it.
For smaller manufacturers across the region — those without the headcount to sustain a dedicated training function or the capital to fund apprenticeship pipelines — this matters in practical terms. The model Autocraft piloted lowers the barrier: skills development becomes embedded in the production process itself, rather than a budget line that competes with operational priorities. In a regional manufacturing context where the gap between knowing and teaching has historically been bridged by proximity and informality, a technology that builds instruction into the task itself is less a disruption than a formalisation of something that was already happening informally.
What a factory in Grantham tells us about how innovation teaches
None of this replaces what a trained engineer knows. AR-guided instruction works at the procedural layer — it can show a worker the correct torque sequence, flag a fitment decision, or walk through a new engine variant step by step. What it cannot do is build the diagnostic reasoning that comes from understanding why a procedure exists, or the broader engineering literacy that underpins a career rather than a task. The model has a ceiling, and it is worth naming it plainly.
Within that ceiling, though, something genuinely instructive is happening in Grantham. Autocraft did not send its workforce on a course. It built the course into the work itself — and in doing so, recast an operational technology investment as a vehicle for continuous knowledge transfer. That is a different definition of innovation from the one most press releases offer: not a product improved, but a relationship between workers and knowledge changed.
The companies most likely to get workforce development right in the next decade may not be the ones that budget the most for training. They may be the ones that stop treating training as a separate event. A factory in Lincolnshire offers a small, precise example of what that looks like in practice.
- [1] Digital twin. https://en.wikipedia.org/?curid=47896295 https://en.wikipedia.org/?curid=47896295
- [2] Fourth Industrial Revolution. https://en.wikipedia.org/?curid=39773873 https://en.wikipedia.org/?curid=39773873