• Written by Muhammed Abdulla NC, CEO

A customer service manager has to decide this week whether to escalate a complaint to the product team, and currently bases it on a mixture of instinct and a brief summary from the support analyst, but this decision could be better informed by applying AI to the customer service data. The manager could use AsscherAi to query the data and get answers in real time, but only if the AI system has been set up correctly. Unfortunately, many organisations make mistakes when applying AI to customer service, which can lead to poor decision making and ineffective use of resources.

Automating a process nobody mapped

One common mistake is automating a process without first mapping out the existing workflow. This can lead to automated systems that do not accurately reflect the way customer service agents actually work. For example, a shift supervisor checks the maintenance log to see which issues have been resolved, but the log does not account for the informal notes and conversations that occur between agents. If the AI system is trained on this incomplete data, it will not be able to provide accurate answers to queries.

A better approach is to take the time to map out the existing workflow, including all the informal processes and exceptions that occur in real-world customer service interactions. This will ensure that the AI system is trained on a complete and accurate picture of the workflow, and can provide reliable answers to queries.

Treating ticket history as complete when it is not

Another mistake is treating the ticket history as a complete and accurate record of all customer interactions. However, ticket history often only captures a fraction of the actual interactions that occur between customers and customer service agents. For example, a customer may have had a conversation with an agent over the phone, but the details of that conversation are not recorded in the ticket history.

This can lead to AI systems that are not able to provide accurate answers to queries, because they are missing critical information about the customer's issue. To avoid this mistake, organisations should take steps to ensure that all customer interactions are captured and recorded, such as using operations management software to track and log all interactions.

Giving the capability only to the analyst who already had it

Some organisations make the mistake of only giving the AI capability to the analyst who already has access to the data and the expertise to interpret it. However, this limits the potential benefits of the AI system, because it does not provide any new capabilities or insights to the rest of the customer service team.

A better approach is to make the AI capability available to all customer service agents, so that they can use it to inform their decisions and provide better service to customers. This can be particularly useful for agents who are dealing with complex or unusual issues, and need to be able to quickly access relevant information and expertise.

Declaring success against a metric with no baseline

Organisations often declare success with their AI-powered customer service systems based on metrics such as response time or resolution rate, but without establishing a baseline for comparison. This makes it difficult to determine whether the AI system is actually having a positive impact, or whether the metrics are simply reflecting normal fluctuations in customer service performance.

To avoid this mistake, organisations should establish clear baselines for their metrics before implementing the AI system, and then track changes in those metrics over time. This will provide a more accurate picture of the AI system's impact, and help organisations to identify areas for improvement.

When AI is not the right choice

While AI can be a powerful tool for customer service, there are some situations where it is not the right choice. For example, in situations where the customer service issue is highly complex or emotionally charged, a human agent may be better equipped to provide the necessary empathy and support. In these cases, organisations should consider using AI to support and augment the work of human agents, rather than replacing them entirely.

Organisations that are considering implementing an AI-powered customer service system should carefully evaluate their needs and goals, and consider whether AI is the right choice for their specific situation. They can contact us to discuss their options and determine the best approach for their organisation.