• Written by Sonali Subhadarsini, Product Manager

This week, the category manager at a large packaging company has to decide whether to switch to a new supplier for a key material, and they are currently basing their decision on a combination of historical data and informal feedback from the production team, which may not be entirely reliable or up-to-date, and could potentially lead to inaccurate quoting and increased substrate waste.

Unique challenges in printing and packaging

The printing and packaging industry presents a unique set of challenges when it comes to implementing enterprise AI governance, particularly in terms of setup and makeready losses, quoting accuracy, and substrate waste, which can all have a significant impact on the bottom line, and are often exacerbated by the job-shop economics that prevail in this sector, where heavy setup costs and frequent changeovers can make it difficult to achieve economies of scale.

For instance, a shift supervisor at a packaging plant may need to balance the need to minimize setup times with the requirement to ensure accurate quoting and reduce substrate waste, all while taking into account the complexities of the production process and the constraints of the equipment and materials being used, which can be a difficult task, especially when the supervisor is not familiar with the specifics of the AI system being used.

Most people get wrong the idea that these challenges can be easily overcome by simply implementing more advanced AI systems, but this is not the case, as the complexities of the printing and packaging industry require a more nuanced approach that takes into account the specific needs and constraints of the sector, and is able to adapt to the unique challenges and requirements of each individual production process.

Enforcing permission at retrieval

Rather than trying to instruct the AI model on what it can and cannot do, which can be a complex and difficult task, it is often more effective to enforce permission at the point of retrieval, where the user is actually accessing the data, which can help to ensure that sensitive information is only available to authorized personnel, and that the AI system is being used in a way that is consistent with the organization's governance policies.

This approach can be particularly useful in the printing and packaging industry, where sensitive information about production processes and materials may need to be protected, and where the consequences of a security breach could be severe, and may involve not only financial losses but also damage to the company's reputation and relationships with its customers.

Auditability

Auditability is a critical component of enterprise AI governance in the printing and packaging industry, as it allows organizations to track what was asked, what was returned, and on what data, which can help to ensure that the AI system is being used in a way that is transparent, accountable, and consistent with the organization's policies and procedures, and can also help to identify potential errors or biases in the system.

For example, a category manager may want to know what data was used to generate a particular quote, or what assumptions were made by the AI model in order to provide a certain recommendation, and being able to access this information can help to build trust in the AI system and ensure that it is being used effectively, and can also help to identify areas where the system may need to be improved or refined.

Routing to a human

There are certain categories of questions or requests that should always be routed to a human by design, rather than being handled by the AI system, such as those that involve sensitive or confidential information, or those that require a level of judgment or expertise that is beyond the capabilities of the AI model, and these categories should be clearly defined and established as part of the organization's governance policies.

For instance, a request to access sensitive information about a competitor's production processes should always be routed to a human, rather than being handled by the AI system, as this information is likely to be confidential and sensitive, and should only be accessed by authorized personnel, and the organization should have clear policies and procedures in place for handling such requests, which can be accessed through the company's intranet.

Limitations and exceptions

While enterprise AI governance can be a powerful tool for improving the efficiency and effectiveness of printing and packaging operations, it is not a solution that is suitable for every organization or every situation, and there may be cases where the costs and complexities of implementing such a system outweigh the potential benefits, or where the organization's specific needs and requirements cannot be met by the AI system, and in such cases, it may be necessary to consider alternative solutions or approaches, such as working with a consultant or expert in the field, who can be contacted through the company's website.