• Written by Azhar Aliyar, Director

A customer service representative checks the ticket history for a recurring issue with a product, only to find that the information is not up to date or is inconsistent with the knowledge base. This is a common problem that arises when AI systems are not properly integrated with the systems of record, leading to a second source of truth that can cause more harm than good.

Read paths, write paths and why the second needs more care

When integrating AI systems with customer service systems, it is essential to consider the read paths and write paths. The read path refers to how the AI system retrieves information from the systems of record, while the write path refers to how the AI system updates or writes new information to these systems. The write path requires more care because it can potentially introduce inconsistencies or errors into the systems of record.

A failure to properly manage the write path can lead to a situation where the AI system becomes a second source of truth, which can cause confusion and inconsistencies. For example, if the AI system updates a customer's order status without properly synchronizing with the order management system, it can lead to discrepancies and errors.

Reconciling definitions that differ between systems

Different systems often have different definitions for the same concept, which can cause problems when integrating AI systems. For instance, the definition of a "customer" may vary between the customer relationship management system and the order management system. Reconciling these definitions is crucial to ensure that the AI system can accurately retrieve and update information.

Most people get wrong the idea that reconciling definitions is a simple task that can be done quickly. However, it requires a thorough understanding of the different systems and their respective definitions, as well as a careful analysis of how these definitions intersect and interact.

The customer service data this depends on

The integration of AI systems with customer service systems depends on several key data sources, including ticket history, knowledge base, product and order records, and escalation history. The ticket history provides a record of all customer interactions, while the knowledge base contains information on products and solutions. The product and order records provide context for the customer's issue, and the escalation history shows how the issue was handled.

A shift supervisor checks the maintenance log to ensure that all customer issues are properly documented and addressed. This requires access to the ticket history and knowledge base, as well as the ability to update the escalation history.

What it changes about

The integration of AI systems with customer service systems changes the way customer issues are handled. Instead of answering the same issue from scratch each time, the AI system can retrieve information from the systems of record and provide a more informed response. Additionally, the AI system can help identify root causes of issues and feed this information back to the product development team.

This can help reduce the number of repeat contacts and improve first contact resolution rates. For more information on how to implement this, see our operations guide.

Measured with

The success of the integration can be measured using key performance indicators such as first contact resolution, time to answer, and repeat contact rate. These metrics provide insight into how well the AI system is performing and where improvements can be made.

By monitoring these metrics, organizations can identify areas where the AI system is struggling and make adjustments as needed. For example, if the repeat contact rate is high, it may indicate that the AI system is not properly retrieving information from the systems of record.

What this does not do

This integration does not replace the need for human customer service representatives. While the AI system can provide automated responses and help with routine issues, complex issues still require the attention of a human representative. Additionally, this integration is not a solution for organizations with poorly designed systems of record or inadequate data quality.

For organizations that are struggling with these issues, it may be more effective to focus on improving their systems and data quality before attempting to integrate AI systems. If you have questions about how to get started, please contact us for more information.