• Written by Azhar Aliyar, Director

A purchasing manager at a manufacturing company wants to know the total value of orders placed with a specific supplier over the last quarter, and they expect to get this information from AsscherAi, the company's new AI-powered query system. To make this work, the data foundations team needs to get started with enterprise data readiness for AI, and the first thirty days are crucial in setting up a system that can provide accurate and reliable answers.

Choosing a first case

When getting started with data foundations, it's essential to choose a first case that already has a known answer to compare against. This approach helps to validate the system and ensures that it's working as expected. In the case of the purchasing manager, the total value of orders placed with a specific supplier over the last quarter is a good starting point because this information is already available in the company's enterprise resource planning system. By comparing the answer provided by AsscherAi with the known answer, the data foundations team can verify that the system is working correctly.

This approach also helps to identify any issues with the data or the system early on, which can save time and effort in the long run. For example, if the answer provided by AsscherAi doesn't match the known answer, the team can investigate the discrepancy and make any necessary adjustments to the system or the data. This iterative process helps to refine the system and ensure that it's providing accurate and reliable answers.

Field-level meaning

One of the common mistakes made when setting up a data foundation system is assuming that column names are sufficient documentation. However, column names often don't provide enough context to understand the field-level meaning of the data. For instance, a column named "order_date" might seem straightforward, but it's not clear whether this refers to the date the order was placed, the date it was shipped, or the date it was delivered.

To avoid this issue, it's essential to document the field-level meaning of the data, including any relevant context or business rules. This documentation should be maintained and updated regularly to ensure that it remains accurate and consistent. By doing so, the data foundations team can ensure that the system is providing answers that are based on a clear understanding of the data.

Systems that get left out of scope

When setting up a data foundation system, it's easy to focus on the most obvious systems and data sources, such as the enterprise resource planning system or the customer relationship management system. However, there are often other systems that get left out of scope, such as legacy systems or departmental databases. These systems may contain critical data that is necessary to provide accurate and complete answers.

For example, the purchasing manager's query about the total value of orders placed with a specific supplier might require data from the accounts payable system, which is not initially included in the scope of the project. If this system is not integrated into the data foundation system, the answer provided by AsscherAi will be incomplete or inaccurate. To avoid this issue, it's essential to identify all relevant systems and data sources and include them in the scope of the project.

What to stop doing if the comparison fails

If the comparison between the answer provided by AsscherAi and the known answer fails, it's essential to stop and investigate the issue rather than trying to force the system to work. This might involve reviewing the data, the system configuration, or the business rules to identify the root cause of the problem.

For instance, the data foundations team might discover that the data is inconsistent or incomplete, or that the system configuration is incorrect. By stopping and investigating the issue, the team can identify the problem and make any necessary adjustments to the system or the data. This approach helps to ensure that the system is providing accurate and reliable answers and prevents the team from wasting time and effort on a flawed system. For more information on troubleshooting, visit our FAQ page.

When it's the wrong choice

While getting started with data foundations is essential for any organization that wants to use AI-powered query systems like AsscherAi, there are situations where it's the wrong choice. For example, if the organization is not ready to invest the time and effort required to set up and maintain a data foundation system, it might be better to delay the project until the organization is more prepared.

Additionally, if the organization is looking for a quick fix or a solution that can be implemented overnight, getting started with data foundations is not the right approach. Setting up a data foundation system requires careful planning, execution, and maintenance, and it's not a project that can be rushed. If you're unsure about whether getting started with data foundations is the right choice for your organization, contact us at our support team for more information.