What data does AI need for production?
A production manager uses AsscherAi to investigate deviations in output against plan by querying the system on cycle times, scrap rates, and maintenance history, then adjusts production schedules and maintenance plans to minimize downtime and optimize production efficiency.
A production manager checks the output against plan for the current shift, comparing the actual production numbers to the scheduled targets, and wonders what data is needed to support an AI system that can provide real-time answers to questions about the production process. The answer lies in the specific systems and fields that a production deployment depends on, including output against plan, cycle times, scrap and quality records, and maintenance history.
The systems that support AI in production
The output against plan is a critical system that provides insights into production efficiency and helps identify areas for improvement. Cycle times, which measure the time it takes to complete a production cycle, are also essential in understanding production capacity and identifying bottlenecks. Scrap and quality records provide valuable information on production quality and help identify areas where processes can be improved. Maintenance history is also crucial in understanding equipment performance and scheduling maintenance to minimize downtime.
These systems are interconnected and provide a comprehensive view of the production process. For example, a shift supervisor checks the maintenance log to ensure that equipment is properly maintained, which in turn affects cycle times and output against plan. The category manager compares two suppliers based on their quality records and delivery times, which impacts scrap and quality records.
What a missing system does to an answer
A missing system can significantly impact the accuracy of an answer provided by an AI system. For instance, if the maintenance history is not available, the AI system may not be able to provide a complete picture of equipment performance, leading to inaccurate predictions of production capacity. Similarly, if scrap and quality records are not available, the AI system may not be able to identify areas for improvement in production quality.
Most people assume that an AI system can still provide accurate answers even with missing data, but this is not the case. The accuracy of an answer depends on the completeness and accuracy of the data, and missing systems can lead to incomplete or inaccurate answers. A production deployment that lacks critical systems can lead to suboptimal decisions and reduced production efficiency.
The definitions that must be agreed
Clear definitions of key terms are essential to ensure that the AI system provides accurate answers. For example, the definition of "production capacity" must be agreed upon to ensure that the AI system provides accurate predictions. Similarly, the definition of "quality" must be agreed upon to ensure that the AI system provides accurate assessments of production quality.
The definitions must be agreed upon by all stakeholders, including production managers, shift supervisors, and category managers. This ensures that everyone is on the same page and that the AI system provides answers that are consistent with the organization's goals and objectives.
Quality problems that only surface later
Some quality problems may not be immediately apparent but can surface later in the production process. For example, a defect in a component may not be detected until the final product is assembled. The AI system must be able to detect these quality problems and provide alerts to production managers and shift supervisors.
The AI system can detect these quality problems by analyzing data from various systems, including scrap and quality records, maintenance history, and output against plan. By analyzing this data, the AI system can identify patterns and trends that may indicate a quality problem. For instance, a sudden increase in scrap rates may indicate a problem with equipment maintenance, which can be addressed by scheduling additional maintenance.
When AI is not the right choice
While an AI system can provide valuable insights and answers to questions about the production process, it is not always the right choice. For example, if the production process is relatively simple and does not involve complex systems and fields, an AI system may not be necessary. In such cases, a simpler solution, such as a spreadsheet or a basic database, may be sufficient.
In other cases, the cost of implementing an AI system may outweigh the benefits. For instance, if the production process is highly variable and requires frequent changes to the AI system, the cost of maintenance and updates may be prohibitive. In such cases, it may be more cost-effective to use a different solution, such as a customized consulting service that provides tailored advice and support.
Frequently asked questions
What data is required to support AI in production?
Output against plan, cycle times, scrap and quality records, and maintenance history are essential systems that support AI in production.
How does a missing system impact AI answers?
A missing system can lead to incomplete or inaccurate answers, as the AI system relies on complete and accurate data to provide reliable insights.
Why are clear definitions of key terms necessary?
Clear definitions ensure that the AI system provides accurate answers consistent with the organization's goals and objectives, and that all stakeholders are on the same page.
When is AI not the right choice for production?
AI may not be necessary for simple production processes or when the cost of implementation and maintenance outweighs the benefits, in which case simpler solutions may be more suitable.