Driving Efficiency: How IT and Engineering Use AI to Boost Manufacturing Productivity
The manufacturing industry is grappling with a pressing challenge: stagnating productivity, particularly in countries with deep industrial roots, such as the United States. Key contributors include insufficient investment, a shortage of skilled labor, high turnover, and the offshoring of production. Ongoing trade disputes exacerbate this issue, as companies scale back investments to offset losses from rising tariffs, especially those dependent on imported components.
Despite the strong pressure for stagnation, the pursuit of continuous improvement and increased productivity is a matter of survival for companies and the economy. In this context, Artificial Intelligence (AI) emerges as a transformative solution to address this paradox, with the IT department playing a pivotal role in identifying solutions aligned with the company’s value chain. The traditional continuous improvement process, supported by AI, will enable companies to shift from a reactive “catch-up mode” to a proactive and predictive approach in such a complex scenario. With reliable data as a foundation, AI can optimize processes, reduce downtime, and enhance continuous improvement. In this article, we explore how the IT department, in partnership with manufacturing engineering, can align data strategies with business objectives to revolutionize productivity in the automotive industry.
Where to Start? Aligning IT and Engineering for AI Projects
Transforming productivity on the factory floor requires close collaboration between manufacturing engineering and the IT department. Together, these teams can identify opportunities for continuous improvement and implement AI-based solutions to address chronic inefficiencies. The starting point is to prioritize processes or workstations with the greatest impact on productivity, using historical data to inform decisions.
1. Identifying Critical Processes by Engineering
Manufacturing engineering must analyze operational metrics to pinpoint bottlenecks. Examples include:
· Downtime: Identifying machines or lines with frequent stoppages.
· Production Efficiency: Comparing designed versus actual parts per cycle.
· Defect Rate: Analyzing the history of rejected parts to detect recurring failures.
· OEE (Overall Equipment Effectiveness): Assessing availability, performance, and quality.
These data-driven analyses enable the prioritization of underperforming processes or production cells, setting clear targets for AI projects.
2. The Role of IT in Data Preparation
Once priority processes are identified, the IT department takes action by mapping the data flow needed to support AI solutions. This mapping occurs even during the ideation phase, before solutions are defined, and involves:
Data Sources: IoT sensors, Manufacturing Execution Systems (MES), ERPs, or manual spreadsheets.
Capture Interfaces: Protocols like MQTT for sensors or REST APIs for MES/ERP integration.
Storage and Structuring: Using data lakes for raw data and data warehouses for processed data.
Initial Governance: Establishing basic policies for data quality and security.
Capturing reliable data in advance is a critical factor in enabling AI adoption, reducing the time between ideation and implementation.
To illustrate, consider an assembly cell producing fewer parts than designed, even after continuous improvement initiatives failed to eliminate chronic issues like unplanned downtime caused by mechanical failures or setup errors. The engineering team, led by an innovation group, analyzes historical data to understand the causes, while IT maps the data flow and interfaces for data capture. A machine learning model trained on vibration and temperature data can predict failures, reducing unplanned downtime. AI-driven solutions, such as intelligent agents, can create and monitor action plans, ensuring all stakeholders execute activities to eliminate the root causes of chronic issues.
The factory floor is a rich and valuable source of data, offering countless possibilities for applying artificial intelligence (AI). From detecting productivity detractors, such as unplanned downtime, to resolving chronic issues like recurring failures in assembly lines, AI can transform operations by leveraging this information. The abundance of data from IoT sensors, MES systems, and other sources enables the formulation of more assertive business cases, with projections of real gains based on precise analyses rather than assumptions or guesses.
Prioritizing initiatives, coupled with robust data governance and the combined expertise of manufacturing engineering and IT teams, forms the fundamental pillar for the success of AI projects on the factory floor. Governance ensures data quality and security, while the technical expertise of both teams enables the identification of high-impact processes and the implementation of predictive and proactive solutions aligned with the company’s strategic objectives.
