" "

How AI Maintains Manufacturing Productivity Amid Reduced CapEx

Manufacturing companies are facing a conundrum: In the face of greater economic uncertainty and rising interest rates, they must maintain high productivity and improved efficiency without increasing CapEx spending. BCG clients have found EBITDA impact as high as +8% through improvements in decision-making related to productivity drivers such as machine setups and changeovers, operator-skill level, unplanned production downtime, and idle time. Recently, many clients have found that they can improve decision making through the use of optimization-based production-scheduling tools.

BCG’s approach to improved productivity

BCG X’s Advanced Production Scheduling (APS) tool transforms what are usually manual, time-consuming approaches to production-scheduling into a rigorous, streamlined process that leverages mathematical optimization to achieve superior scheduling results. After just eight weeks of using APS, our clients have seen a more than 3% OEE uplift (equivalent to approximately 30 minutes of additional production time per day) by changeover reduction and asset-capacity improvement, and a greater than 50% reduction in planning-related labor hours for scheduling, along with more sustainable operations, greater customer-service levels, and more employee engagement.

APS addresses four common pain points: manual processes, conflicting objectives, disparate data sources, and analytical and computational capacity.

First, it takes cumbersome, non-standardized decision-making processes often built in spreadsheets and replaces them with data-based, algorithm-driven processes.

Second, it uses advanced data science to resolve often conflicting objectives such as production efficiency versus delivery commitment/speed to delivery, cost versus customer satisfaction, and flexibility versus higher cost and lower speed.

Third, it gathers data across disparate sources, using an AI-driven text-mining tool when necessary to clear the data for consistency, and then consolidates the data into a structured, usable form in a single data lake.

Finally, it provides the mathematical models and the computing capability needed to effectively plan scheduling to make intelligent decisions.

The AI-based Approach: Achieving operational goals through scheduling grounded in data science

The APS scheduling algorithm synthesizes datasets into an optimal production plan using mixed-integer programming, simulations, and heuristics. The tool weighs multiple objectives to create optimal schedules that factor in an organization's diverse operational goals. The multiple-objective design reflects the fact that an enterprise can have more than one operational goal. Such goals might include:

  • Changeover minimization: Maximizing the utilization time of machines to reduce average production costs
  • On-time in-full delivery rate: Maximizing customer satisfaction by meeting different OTIF rates for different customer tiers
  • Machine backlog balance: Making sure the configured resources (machine and operator) have relatively even amounts of work to do
  • Cross-step backlog balance: Avoiding the need to wait for the arrival of output from previous manufacturing stages or to clear long queues, either of which would require the current processing step to alter the number of units to be produced in the near future

In APS, operational situations are modeled as constraints, with the algorithm using multiple mathematical programming techniques to find optimal answers in a reasonable time. Typical operational constraints include:

  • Machine unavailability due to maintenance, operator absence, etc.
  • Unavailability of raw materials
  • Lack of machine-product compatibility (whether a product can be produced by a certain type of machine)
  • Throughput capability of machines based on configuration
  • RM check and allocation
  • Outsourcing

This mathematical construct is supplemented by an ML-driven engine that detects changes based on new information (e.g., urgent orders coming in, critical materials running out, prior manufacturing stages not yielding desired level of throughput) and generating a next set of actions for planners and schedulers.

The underlying algorithm processes data through a decision framework to create an optimal schedule

Implementing APS

When working with companies to implement new technologies, BCG follows the 10-20-70 formula for dividing AI project investments:

  • 10%: AI and machine-learning algorithms and optimization, and data analytics
  • 20%: Software and technology infrastructure to optimize, integrate, and scale the AI and machine learning
  • 70%: Consulting, domain expertise, project management, and change management to internalize the AI transformation and tie all projects back to established business goals and objectives

The key to successful APS adoption is the proper allocation of resources to fully integrate the tool into business processes across the organization. Doing so helps unlock long-term value by addressing key pain points and questions for stakeholders across functions, while enabling teams to plan more proactively and accurately.

Organizations often cite integration as a key difficulty in their digital transformations. In surveys conducted by BCG, both managers and employees note that the biggest issue they face in digitally transforming a process within an organization is building the right people skillsets. To address this, BCG works alongside organizations to enable their employees to adopt and integrate the new digital process.

BCG also supports skill building and knowledge transfer so that organizations can scale the APS approach long after its engagement with BCG has ended. Using an agile development process, teams are involved in design discussions to provide input and are kept abreast of progress so that they feel empowered to take ownership of the final product.

Additionally, BCG establishes a thorough handoff process that includes creation of a client-side “digital squad” whose members acquire project ownership over a phased implementation. BCG also provides hands-on knowledge transfer sessions and detailed documentation. The same digital squads can be tasked to build subsequent digital use cases.

People & Processes: Thorough handoff process supports ongoing implementation

Advanced Production Scheduling enables companies to maintain—and often increase—their manufacturing production goals even as CapEx budgets are trimmed in response to macroeconomic conditions. BCG recommends APS to manufacturers as part of the broader transformation of their business processes. Effective APS integration will help companies unlock the technology’s true value as it brings the power of advanced analytics to more and more of their traditionally manual processes.

How AI Maintains Manufacturing Productivity Amid Reduced CapEx