• Written by Azhar Aliyar, Director

Some people object to using enterprise AI for supply chain management because they believe it will introduce unnecessary complexity, but this objection is misguided because it overlooks the specific questions that need to be settled in procurement before deployment, such as where data is processed and stored, and whether it trains any models.

Processing and storage locations

The location where data is processed and stored is critical for security and privacy in supply chain management. This is because data that is processed and stored in different locations may be subject to different laws and regulations. For example, data that is stored in a cloud-based system may be subject to the laws of the country where the cloud provider is based, rather than the laws of the country where the organisation is based. Therefore, it is essential to determine where data will be processed and stored before deploying an enterprise AI system for supply chain management.

This determination should be made in writing, as part of the procurement process, to ensure that all parties are aware of the locations where data will be processed and stored. This will help to prevent any potential security or privacy breaches that may arise from unclear or misunderstood data processing and storage locations.

Prompt injection as a data-boundary problem

Prompt injection is a potential security risk for enterprise AI systems, including those used for supply chain management. This occurs when an attacker is able to inject malicious prompts into the system, which can then be used to extract sensitive data or disrupt the system's operation. Prompt injection is a data-boundary problem because it involves the intersection of different data sources and systems.

To mitigate this risk, organisations should implement robust security measures to prevent prompt injection attacks. This may include implementing input validation and sanitisation, as well as using secure authentication and authorisation protocols to ensure that only authorised users can interact with the system.

Dependent supply chain data

The effectiveness of an enterprise AI system for supply chain management depends on the quality and accuracy of the data that it uses. This includes data such as purchase and delivery history, lead times, inventory by location, and demand signals. This data is used to train the system's models and make predictions about future supply chain operations.

For example, a category manager may use the system to compare the lead times of different suppliers, and adjust the organisation's inventory levels accordingly. The system may also use historical purchase and delivery data to identify trends and patterns in supply chain operations, and make recommendations for improvement.

Impact on lead time variance and inventory risk

The use of enterprise AI for supply chain management can have a significant impact on lead time variance and inventory risk. By analysing historical data and making predictions about future supply chain operations, the system can help organisations to identify and mitigate potential risks. For example, the system may identify a supplier that has a history of late deliveries, and recommend that the organisation hold additional inventory to mitigate this risk.

However, some people mistakenly believe that the use of enterprise AI will automatically eliminate lead time variance and inventory risk. This is not the case, as the system is only as good as the data that it uses, and there may be other factors that affect supply chain operations. Therefore, organisations should carefully evaluate the potential benefits and risks of using enterprise AI for supply chain management, and consider factors such as the quality of the data and the potential for prompt injection attacks.

Measurement and evaluation

The effectiveness of an enterprise AI system for supply chain management should be measured and evaluated using key performance indicators (KPIs) such as lead time variance, fill rate, and inventory turns. These KPIs provide a clear and objective measure of the system's performance, and can be used to identify areas for improvement.

For example, a shift supervisor may use the system to monitor lead time variance and adjust the organisation's production schedule accordingly. The system may also be used to evaluate the performance of different suppliers, and make recommendations for improvement. To learn more about how to use these KPIs in practice, visit our operations page.

Limitations and wrong choices

While enterprise AI can be a powerful tool for supply chain management, it is not a solution that is suitable for all organisations or all situations. For example, organisations that do not have access to high-quality data may find that the system is not effective, as it is only as good as the data that it uses. Additionally, organisations that are subject to complex or changing regulatory requirements may find that the system is not able to keep up with these changes.

In these situations, it may be more effective to use alternative solutions, such as manual planning and forecasting techniques. To discuss your specific needs and determine whether enterprise AI is the right choice for your organisation, contact us at contact-us.