Customers expect immediate answer to their concern
With the rise of social media and messaging, today’s customers have a much higher expectation of customer service than they did a few years ago. They expect an immediate answer to their concern and are no longer willing to wait. Today, retail companies receive hundreds of e-mail enquiries and questions every day. Assigning them isn’t as simple as it seems. Sending enquiries to the right person is time consuming and error-prone. Also, with the increasing number of e-mails, the quality and rapidity decreases – and therefore customer satisfaction goes down the drain. Those requests then have to be manually sorted into one of more than 20 categories, then treated by the agent specialized in the related domain. This again is prone to consume time and makes mistakes.
Artificial Intelligence can speed up the process
So how can our retail client speed up the process while also making it more accurate? The answer is artificial intelligence. OpenWT was asked to provide an artificial intelligence system which would classify the different e-mails and directly send them to the mailboxes of the responsible agents, while also supporting the agent in answering the requests to improve drastically the quality of the service.
How we integrated our AI algorithm into their system
- Identification of the elements which impact the classification process.
- Establishing of a list of contextual information such as language identification and enriched customer information as well as a list of desired standard answers.
- Training of the AI on feature engineering and natural language processing so it understands the emails.
- Training of the AI on e-mail classification to make it very stable and highly performing.
- Training of the AI on e-mail rendering to allow it to combine name entity recognition with the CRM to know the client and propose an answer to the specialized agent.
The AI can analyze the email received by customer support. It uses the data collected to gather information about the sender, the topic and the problem mentioned.
Once the AI has analyzed the content of a request, it can process it and send it to the agent specialized on the right topic, reducing time and errors.
More than just attributing the email to the right mailbox, the AI can prepare an answer that the specialist will just have to validate, which is, again, drastically reducing time, effort and errors.
Impact by leveraging AI
hit rate (better than humans)
of cost reduction
of proven accuracy