• Written by Muhammed Abdulla NC, CEO

A production line grinds to a halt because a critical machine is waiting for a part that was supposed to be delivered the previous day, and the team leader is left scrambling to explain why the delivery was delayed, only to discover that the issue was caused by a combination of factors that were not immediately apparent, including a miscommunication between the supplier and the logistics team, and a software glitch that failed to alert the team to the potential delay. This kind of anomaly, which is only visible when looking at the combination of multiple factors, can be particularly challenging to identify and resolve. The question is, can AI help reduce these kinds of anomalies?

What has to be in the data for this to be possible

To use AI to reduce anomalies only visible in combination, the data must include detailed records of processes and exceptions, as well as project status updates. This means that the organisation must have a robust system in place for tracking and recording data, such as a operations management system. The data must also be accurate and up-to-date, and must include information on all the relevant factors that could contribute to anomalies, such as supplier performance, logistics, and software glitches.

A shift supervisor, for example, might check the maintenance log to see if there were any issues with the machine that could have contributed to the delay, while also reviewing the project status updates to see if there were any changes or exceptions that could have affected the delivery. By looking at the combination of these factors, the supervisor can start to identify potential causes of the anomaly.

The mechanism, stated plainly

The mechanism by which AI can help reduce anomalies only visible in combination is through the use of algorithms that can analyse large datasets and identify patterns and correlations that may not be immediately apparent. These algorithms can be trained on historical data to learn what normal patterns of behaviour look like, and can then be used to identify anomalies and alert the team to potential issues.

For example, an algorithm might be trained on data from previous deliveries to learn what factors contribute to delays, and can then be used to analyse data from current deliveries to identify potential risks. By using AI in this way, organisations can proactively identify and address anomalies, rather than simply reacting to them after they have occurred.

What it will not fix

While AI can be a powerful tool for reducing anomalies only visible in combination, it is not a silver bullet. There are many cases where AI will not be able to fix the issue, such as when the anomaly is caused by a factor that is not included in the data, or when the data is incomplete or inaccurate. Most people assume that AI can automatically identify and fix all anomalies, but this is wrong - AI is only as good as the data it is trained on, and can only identify patterns and correlations that are present in that data.

In some cases, the anomaly may be caused by a complex interplay of factors that is difficult for AI to identify, or may require a level of human judgement and expertise that is difficult to replicate with algorithms. In these cases, AI may be able to provide some insights and suggestions, but will not be able to fully resolve the issue.

How you would know it worked

One way to know if AI is effectively reducing anomalies only visible in combination is to look at on-time delivery rates. If the organisation is able to deliver products on time more consistently, then it is likely that the AI is having a positive impact. A category manager, for example, might compare the on-time delivery rates for different suppliers, and use this information to identify areas where the AI is having the most impact.

By tracking on-time delivery rates and other key performance indicators, organisations can get a sense of whether the AI is having the desired effect, and can make adjustments as needed to optimise its performance. This might involve refining the algorithms, updating the data, or providing additional training to the team.

What this does not do

This approach to using AI to reduce anomalies only visible in combination does not replace the need for human judgement and expertise. While AI can provide valuable insights and suggestions, it is still important to have a team of experienced professionals who can review and interpret the results, and make decisions based on their expertise. If you are interested in learning more about how AsscherAi can help your organisation reduce anomalies, contact us to discuss your needs and find out how we can help.