• Written by Azhar Aliyar, Director

A production planner at a tier one supplier checks the inventory levels of a critical component, only to find that the numbers do not match the expected usage rates. This discrepancy can have far-reaching consequences, from delayed shipments to costly rework, and highlights the importance of reliable data in the automotive supply chain. The ability of AsscherAi to answer questions about an organisation's own data in real time relies on a foundation of accurate and consistent information, which can be particularly challenging to achieve in the automotive industry.

Complexity in automotive data

The automotive industry presents unique challenges when it comes to data management, particularly in the context of tiered supply chains with tight takt times and long programmes. Supplier escalations, for example, can lead to a cascade of changes that must be propagated throughout the supply chain, making it difficult to maintain data consistency. Engineering change propagation is another area where data accuracy is crucial, as even small changes can have significant impacts on production and quality. Warranty root cause analysis also relies on accurate data, as incorrect or incomplete information can lead to misidentification of root causes and ineffective corrective actions.

Most people get wrong the idea that these challenges can be overcome simply by implementing more advanced technology or relying on manual data entry. However, the reality is that these approaches often introduce new errors or inconsistencies, and can even exacerbate existing problems. A more effective approach is to focus on establishing robust data foundations, including clear definitions, standardized processes, and automated data validation.

Field-level meaning and documentation

One common misconception is that column names in a database are sufficient documentation for the data they contain. However, this is rarely the case, as column names often lack the context and nuance necessary to fully understand the data. Field-level meaning, on the other hand, requires a deep understanding of the specific business processes and rules that govern the data. This includes knowledge of data formats, validation rules, and relationships between different fields and tables.

A good example of this can be seen in the purchase order table, where a column named "status" might seem straightforward, but actually requires knowledge of the specific workflow and business rules that govern the status updates. Without this context, it is easy to misinterpret the data or make incorrect assumptions about its meaning.

Systems and scope

In many cases, certain systems or data sources are left out of scope when implementing AsscherAi, which can lead to partial answers or incomplete information. This might include legacy systems, external data sources, or even manual processes that are not fully integrated into the overall data landscape. As a result, the answers provided by AsscherAi may not reflect the full complexity of the business, leading to decisions that are based on incomplete or inaccurate information.

For instance, a category manager comparing two suppliers might not have access to data from the enterprise resource planning system, which could provide critical information about lead times, inventory levels, and quality metrics. Without this information, the comparison is incomplete, and the decision may not be optimal.

To avoid this, it is essential to carefully consider the scope of the implementation and ensure that all relevant systems and data sources are included. This might involve working with stakeholders to identify key data sources, integrating with external systems, or developing custom connectors to legacy applications. By taking a comprehensive approach, organisations can ensure that AsscherAi provides accurate and complete answers to their questions. For more information on implementing AsscherAi, visit our website.

Gaps and biases in the data

Stockouts, backfills, and other gaps in the data can also teach a model the wrong lesson, leading to inaccurate predictions or recommendations. This is because the model is trained on incomplete or biased data, which can perpetuate existing errors or inconsistencies. For example, if a model is trained on data that is biased towards a particular supplier or region, it may not generalize well to other scenarios or contexts.

A worked example of this can be seen in the case of a shift supervisor checking the maintenance log, where gaps in the data might lead to incorrect conclusions about equipment reliability or maintenance schedules. By identifying and addressing these gaps, organisations can ensure that their models are trained on accurate and representative data, leading to better decision-making and more effective operations.

Limitations and alternatives

While AsscherAi can provide significant benefits in terms of data analysis and decision-making, it is not a solution for every problem. In some cases, the data may be too complex or nuanced to be effectively addressed by a machine learning model, or the business processes may be too dynamic or rapidly changing. In these situations, alternative approaches such as manual analysis or expert judgment may be more effective. If you are unsure about whether AsscherAi is the right choice for your organisation, contact us to discuss your specific needs and challenges.