The ask
With Ticket Next Best Action (TNBA) in full production, telecommunications company MetTel was looking for ways to elevate customer service levels even further. TNBA was already adding considerable value to the maintenance ticket process, so the company was curious if the same A.I. enabled process automation technology could present a new opportunity at a different customer touchpoint.
In their existing resolution process, each MetTel customer used a unique email address to send a range of requests, including scheduling a repair or service, changing an order, placing a new order, or asking a billing question. MetTel’s team would then manually read and classify these emails, extracting the information, and raising them as tickets. For maintenance requests, these tickets would be handled by TNBA, but for all others, manual processes would be used to resolve the issues.
We identified the email classification and ticket raising activity as a prime candidate for A.I. enabled process automation, due to the volume and structured nature of the company’s existing historical data. Repair requests were the highest type of email by volume, and so this is where we began.
Key objective: Elevate customer service levels by extending the chain of automation to email requests.
What we did
We designed and deployed a system capable of classifying each email, intelligently extracting it’s key content, and automatically generating an appropriate ticket.
A.I. is used to sift through the content of the email, assign its category, and extract relevant data required to populate the fields of a ticket. For example, a billing request will need information such as the customer company name, product they are using, pricing plan etc.
We also built an additional function should the customer email not contain this required information. A virtual assistant automatically responds to the email requesting any missing information that is required for the ticket. Customers can quickly respond, and the email can be automatically classified.
Once an email is classified, its generated ticket is sent to the right team, which can act accordingly, track progress, and ensure SLAs are being met. This saves a welcomed amount of time for MetTel’s teams, but the solution unlocks its maximum value for repair requests, thanks to its seamless integration with TNBA.
Now, when a repair ticket is generated by the email system, it is automatically entered into TNBA, where automation continues throughout resolution. For certain cases, this means full automation from the initial customer email request, all the way to the issue being resolved.
The process
Step One
Having already completed the TNBA project, we already had a strong understanding of the elements needed for maintenance ticket creation. In a similar way to how tickets are classified, we needed to build the functionality of classifying emails before ticket creation could occur. To achieve this, first we aggregated data from historical emails to better understand the different ways they could be classified, and determined what constitutes a repair request.
Step two
We then built a model using Natural Language Processing (NLP) capable of reading and understanding the content of the emails so that they can be accurately classified. This model was trained on the historical data collated in step one.
Step three
Lastly, we defined the required integrations to enable the email classification solution to work seamlessly with TNBA for instances of repair requests. This includes ensuring the required ticket fields are defined and complete, and the ticket is automatically entered into the system where it’s journey can continue.
Result
With the email classification system in place, 98% of all repair request emails are now accurately classified, and 68% of classified emails can have tickets automatically created and entered into TNBA. MetTel is now able to process thousands of emails per month without any human interaction. This has not only unlocked hours of time, but the speed of response and adherence to SLAs has massively increased, helping the company achieve its original goal of improved service. By removing the manual handling of tickets, the solution also enables infinite scaling, thanks to its capacity to manage an unlimited number of repair requests – even as MetTel’s customer base continues to grow.
