• Written by Azhar Aliyar, Director

A finance manager stares at a dashboard, wondering why the new enterprise AI system is not being used to its full potential, despite the significant investment made in its implementation. The system is capable of answering complex questions about the organisation's data, but it seems to be gathering dust.

Attaching new capability to a routine that already exists

The problem often lies in the fact that new systems are introduced without being properly integrated into existing workflows. For instance, a shift supervisor checks the maintenance log every morning, but the new AI system is not linked to this routine, making it easy to forget or ignore. To change this, the AI system needs to be attached to a routine that already exists, making it a natural part of the workflow.

This can be achieved by identifying the key tasks and routines that are already in place and finding ways to incorporate the AI system into these processes. For example, the finance manager could set up a daily report that uses the AI system to provide insights into the organisation's financial data, which would then be discussed in the daily meeting.

Who needs to be able to ask, not just who needs the answer

Most people get wrong the idea that only a select few need to be able to ask questions of the AI system, when in fact, it is often the case that many people need to be able to ask questions in order to get the most out of the system. The category manager, for instance, may need to compare two suppliers, and the AI system can provide the necessary data to inform this decision.

By allowing more people to ask questions, the organisation can unlock the full potential of the AI system and ensure that it is being used to its full capacity. This can be achieved by providing training and support to a wider range of users, rather than just a select few.

The first wrong answer and how trust is rebuilt

The first time the AI system provides a wrong answer, it can be a significant setback, as users may lose trust in the system. However, this is not necessarily a reason to abandon the system, but rather an opportunity to rebuild trust. The finance manager, for instance, may need to investigate why the wrong answer was provided and take steps to correct it.

By being transparent about the limitations of the AI system and taking steps to correct mistakes, the organisation can rebuild trust with its users. This can involve providing clear explanations of how the system works and what it can and cannot do, as well as providing ongoing support and training to users.

Training that is about the domain, not the interface

Training for the AI system is often focused on the interface itself, rather than the domain in which it is being used. However, this approach is wrong, as it does not provide users with the necessary context and understanding of how to use the system effectively. The category manager, for instance, needs to understand how to use the AI system to inform procurement decisions, rather than just how to use the interface.

By providing training that is focused on the domain, rather than the interface, the organisation can ensure that users have the necessary skills and knowledge to use the AI system effectively. This can involve providing training on the key concepts and terminology of the domain, as well as how to apply the insights provided by the AI system.

What this does not do

The AI system is not a replacement for human judgement and expertise, but rather a tool to support and inform decision-making. It is not a solution to all problems, but rather a solution to specific problems that involve complex data analysis. For instance, the finance manager may use the AI system to provide insights into the organisation's financial data, but ultimately, the decision of what to do with that information rests with the manager.

By understanding what the AI system can and cannot do, the organisation can ensure that it is being used effectively and that users have realistic expectations of what it can achieve. If you are interested in learning more about how to implement an AI system in your organisation, you can visit our website at www.asscher.ai for more information.

Common pitfalls to avoid

One of the common pitfalls to avoid when implementing an AI system is to assume that it will be widely adopted without any effort or support. However, this is often not the case, and significant effort and resources are required to ensure that the system is being used to its full potential. The organisation needs to be willing to invest time and money in training and supporting users, as well as in maintaining and updating the system.

By being aware of these common pitfalls, the organisation can take steps to avoid them and ensure that the AI system is being used effectively. If you have any questions or need further guidance, you can contact us at contact-us@asscher.ai for more information.

Conclusion of the process

The process of implementing an AI system is not a one-time event, but rather an ongoing process that requires continuous effort and support. The organisation needs to be willing to invest time and money in maintaining and updating the system, as well as in training and supporting users.