Status quo, trends and challenges
How far can vehicle assembly automation really go?
The automation of vehicle assembly is gaining importance. Flexible lines, AI-supported processes and smart tooling help OEMs to manage variant diversity, cost pressure and skills shortages. An overview.
The production of a vehicle follows a clear logic, even in the interplay of automation and human labour. In the press shop and body shop, highly automated processes dominate, where employees mainly take on monitoring, control, and maintenance tasks. The paintshop is also largely robotised today. People are mostly involved in peripheral processes such as preparation, quality assurance, or special cases.
With the transition to assembly, this picture changes fundamentally. The closer the vehicle comes to its final configuration, the greater the human share in the processes. The diversity of variants, manual joining tasks, quality decisions, and ergonomically demanding activities make assembly the most labour-intensive area of vehicle production to this day.
At the same time, assembly is the area where automation has long progressed at the slowest rate. This was not due to a lack of technology, but framework conditions: frequent product changes, complex spatial references, a close dependence on material supply and quality assurance, as well as the immediate effects of deviations on cycle, throughput time, and output.
This structural restraint is also reflected in the degree of automation. "The degree of automation in final assembly in Western Europe today is between 20 and 35 per cent, depending on the plant and product, and is therefore significantly below the level of body construction and painting," explains Dr. Werner Geiger, managing director of German management consultancy, Agamus Consult. Within final assembly, areas with heavy or highly standardised activities are primarily automated, such as the marriage of the powertrain, test benches and tests, as well as glazing.
This situation has intensified in recent years. Cost pressure is increasing, while volume and model mix are becoming increasingly difficult to plan. At the same time, the shortage of skilled workers is worsening. Naturally, shift work and physically demanding activities are not everyone's cup of tea. In addition, there is the parallel production of different drive concepts. Combustion engines, hybrid and electric vehicles are increasingly running on the same assembly lines, but with different space concepts, process sequences and quality requirements.
The need for cost reduction in Western automotive assembly factories has been a roadblock to the increase of automation. These factories often operate in overcapacity and therefore do not see an immediate need for more automation.
From a strategic point of view, cost pressure is the main factor slowing down the further expansion of assembly automation in Western Europe, adds Pedro Pacheco, vice president analyst at global research and advisory firm, Gartner. Many plants are already operating with overcapacity and therefore see no immediate economic imperative to install additional automation. The situation is developing much more dynamically in China and among new market entrants and individual greenfield projects of established OEMs. There, automation is understood as a strategic future issue and is being consistently pursued accordingly.
Automation is understood in this context, less as a means of achieving maximum efficiency, but rather as a tool for stabilising and controlling assembly. OEMs and suppliers are relying on solutions that can be integrated into existing lines, map variant changes via software, and be gradually expanded. The actual degree of automation in final assembly remains comparatively low and has so far increased incrementally rather than in leaps and bounds, especially compared to body construction and painting. Assembly automation is thus evolving from a one-off measure to a structural design principle.
Flexible assembly lines without commitment
A central goal of modern assembly automation is to decouple production lines from a fixed product logic. Instead of designing lines for individual models or drive types, processes are designed to remain flexible within a defined range of variants. In practice, however, this approach reaches its limits where the classic assembly line principle makes it difficult to maintain stable control of OEE (Overall Equipment Effectiveness) and disruptions to automated stations have immediate effects on large parts of the line.
Automation can be specifically applied where it improves repeatability, process stability, and ergonomics. Where variance unnecessarily complicates, it brings no added value.
Flexible and modular automation concepts reduce investment risk, but fully automated stations also increase systemic risk, as a single failure can affect large parts of the assembly line and overall OEE.
A striking example is the automated marriage at the VW plant in Emden. The joining process between the body and the powertrain is fully automated and electronically controlled there. The system recognises the incoming vehicle and automatically adjusts gripping points, movement profiles, and joining strategies. Different chassis and battery variants can thus be processed on the same line without manual conversion.
Such applications represent a fundamental change. Instead of continuous full automation, modular automation stations with clearly defined tasks are emerging. These islands can be integrated into existing lines, piloted, and expanded or reconfigured as needed. For OEMs, this means more investment security in an environment of uncertain sales forecasts.
“Drivers of this development are primarily flexible, modular automation concepts that can be better secured through digital commissioning, simulation, and digital twins, thereby reducing investment risk,” explains Geiger. At the same time, system availability plays a central role. “If a fully automated station fails, not only one process but large parts of the line, including the employees, are at risk.”
Process design is also crucial in this context. Automation-friendly assembly requires clear positional references, defined joining points, and a deliberate bundling of variants. The earlier product and process design are geared towards automation, the lower the integration effort later on. Geiger sees limitations where cost-benefit ratios do not add up, the variety of variants is too high, or sensitive tasks are required. The classic assembly line principle also makes it difficult to control OEE in many cases.
AI-supported sequencing and decision logic at the station
With increasing variety of variants, the complexity at individual assembly workstations rises. Which parts are needed when? In what order should work steps be carried out to optimise cycle time, quality, and ergonomics? Classical, statically planned sequences reach their limits here.
Artificial intelligence is therefore gaining importance in assembly, less as an autonomous system and more as data-based decision logic. AI-supported applications analyse process, quality, and tool data in real-time and directly support process control at the station. Decisions about sequences, rework, or skipping individual steps can thus be made situationally.
Even the most advanced factories in the world are not fully automated. Areas such as cabin assembly and wire harness installation remain the final frontier, as they are so challenging that they can still only be done by human workers in most cases.
One area of application is AI-supported quality inspection in assembly. BMW uses AI in several plants to detect deviations early and evaluate them automatically. The decision on whether a vehicle continues or goes for rework directly influences the subsequent process sequence.
“AI helps in assembly today, especially where processes are stable and high-quality data is available, such as in camera-based quality inspection or the evaluation of screw data,” says Geiger. In practice, it is mainly about support, such as recognising and prioritising anomalies faster, not about AI independently controlling the assembly.
From a strategic perspective, Pacheco also warns against exaggerated expectations. Autonomous machines, robots, or AGVs controlled in real-time by a predictive system are a possible future vision but are often overestimated today. Without standardised processes and clean data, AI is often expected to perform miracles, “even though you have to learn to walk before you can run.”
A prerequisite for meaningful AI use remains a reliable data basis. Tool curves, image and sensor data, process times, and material IDs must be cleanly recorded and traceable. Governance is equally important. Which decisions can the system make autonomously, and which require human approval? In assembly, it is not the complexity of the model that decides, but the stability of the result.
Cost-efficient automation instead of large investments
In addition to technological development, the economic approach to assembly automation is also changing. Large projects with high investment volumes and long payback periods are increasingly less suited to volatile markets. Many OEMs and suppliers are therefore relying on incremental automation. This approach particularly shapes established plants in Europe and North America, while new market entrants and several Chinese OEMs are consciously focusing much more on comprehensive automation.
Individual processes are automated, tested during ongoing operations, and scaled if successful. Collaborative robots, autonomous mobile robots, and standardised automation modules enable a comparatively low entry point. Automation thus becomes a continuous development process rather than a one-time conversion.
Successful assembly automation requires all key levers to be addressed at the same time. Technology, process design and workforce qualification must be developed together, as isolated optimisation rarely leads to a stable overall system.
“We clearly observe a shift away from large, highly specialised automation projects towards modular and incremental solutions,” confirms Geiger. Drivers include the increasing volatility of markets, the risk of sunk costs, and the better ability to reuse plant concepts across model generations. Area productivity and risk minimisation also play an increasing role, as disruptions in highly automated lines can quickly propagate.
Ford consistently follows a brownfield approach in assembly, for example, and gradually integrates automation into existing lines, such as through mobile robotics and modular assistance systems. This pragmatic approach is also evident among suppliers. ContiTech has set up an automated production cell at its Hannover-Vahrenwald plant, which combines several process steps while deliberately retaining manual visual and dimensional inspection.
However, Pacheco points out that this trend is not uniform globally. While financial constraints and overcapacity led to caution among many established OEMs in Europe and North America, new market entrants and numerous Chinese manufacturers continue to consistently focus on automation. Gartner expects that at least one OEM will realise a fully automated assembly line before the end of this decade.
Smart tooling as an enabler for variant diversity
In addition to robotics and IT, the tool itself is increasingly coming into focus. Smart tooling automatically recognises which vehicle or variant is being processed and independently adjusts parameters such as torque, programmes, or positions. This reduces errors and shortens changeovers.
In the Volkswagen plant in Emden, height-adjustable hangers are used, which automatically adjust their working height to the respective vehicle and communicate with the line control via transponders. At Forvia Hella, smart tooling is also part of an overarching production system. Tool, process, and quality data come together, enabling quick, data-based decisions in assembly.
“Areas such as powertrain assembly, marriage, quality inspection, and vehicle commissioning, as well as activities with the same take rate, such as mounting wheels, seats, or glazing, are particularly suitable for automation today,” says Geiger. Ergonomically demanding tasks are also predestined for automation. At the same time, areas that have been weakly automated so far, such as the wiring harness, are increasing proportionally, not least due to changed product architectures like distributed assembly.
Humans and automation in assembly
Even though the point could hardly seem more clichéd, humans are indeed at the centre of assembly despite increasing automation. However, their role is shifting. Instead of repetitive, ergonomically demanding tasks, workers are increasingly taking on monitoring, control, and problem-solving tasks. Automation can relieve and stabilise, not displace.
Of fundamental importance is the acceptance on the shop floor. Systems must be transparent, provide clear benefits, and integrate seamlessly into existing processes. Especially with AI-supported decisions, traceability is central. Responsibility must not become blurred.
The role of data and AI is very pervasive, but expectations often exceed reality. Without strong data foundations and process standards, companies expect AI to deliver disruptive improvements that it is not yet able to achieve.
From a strategic perspective, additional factors come into play, emphasises Pacheco. Companies must be culturally ready to allow radical changes in manufacturing, while also dealing with established IT and OT structures. Additionally, vehicle design is a central lever. Those who want to approach 100% automation must design products so that they can be assembled robotically. Especially in Europe, the social and political dimension also plays a role. "Companies need convincing concepts on how increasing automation is not automatically equated with job loss."
Many plants are responding with new organisational models. Interdisciplinary teams from production, maintenance, IT, and engineering are iteratively developing automation solutions. "New automation requires new role models," says Geiger. He cites as an example employees from assembly or injection moulding who are qualified as mobile plant operators and now oversee entire fleets of AGVs.
Assembly automation as system design
The automation of vehicle assembly is evolving from a piecemeal measure to a holistic system design. Flexible automation stations, AI-supported decision logic, cost-efficient robotics, and smart tooling make it possible to gradually stabilise assembly processes and adapt to new variants.
“Just as a car needs four wheels, successful assembly automation requires all key levers simultaneously,” summarises Geiger. Technology, process design, qualification, and organisation must be considered together. A holistic view is necessary in all phases of plan, build, and run, as the sum of individual optima has never been the overall optimum.
The balance remains crucial. Standardisation where it brings quality and cycle stability, and flexibility where variance is unavoidable. Assembly automation thus becomes a central lever for competitiveness, not because it is maximally automated, but because it makes assembly manageable in a volatile environment.
