Serving 350M+ customers better: Leveraging data to power complaint resolution
Superior customer service entails being responsive to a customer’s needs throughout the customer lifecycle. From a cold lead to an acquired customer and beyond. Reports suggest that most customers are willing to pay a premium for better experience. Beyond financial performance, this helps in higher customer retention, which is significant for a high-churn industry such as Telco. With a constantly expanding user-base and evolving product offerings, we, at Airtel continuously strive to deliver exceptional customer experience. Streamlined complaint resolution process plays a vital role in making us stand out. At the scale of more than 350 Mn customers, it becomes even more critical.
Point of departure
Customer complaint resolution presents a set of fundamental challenges to the business. These include, low levels of automation, scope of human error, high resolution time, and inefficient call centre processes. Managing grievances becomes even more challenging with our large scale and continuously changing expectations. We receive millions of network complaints every month. As a result, earlier, we often missed SLAs which led to poor customer experience. Primary root cause of these issues was the lack of an intelligence layer to support problem identification. Having identified an exhaustive set of challenges and a clear objective of improving complaint resolution, we created X-Net as a solution. X-Net is an in-house Network Complaint Handling engine to identify and classify grievances efficiently.
A closer look at X-Net
We receive customer complaints from a variety of channels which include call centres, IVR, Thanks app, and others, such as social media, and email. With the basic details that a customer provides, it is difficult to identify the root of the problem. X-Net helps us zoom into the intricate network infrastructure and find the epicentre, colloquially called the problematic cell. X-Net enables this with accuracy, speed and without manual intervention. Finding the problem accurately takes a sizeable portion of total resolution time. Therefore, faster identification of the problem considerably improves our resolution cycles.
The classification engine in X-Net starts by collating most used cells (towers) by the complainant. These cells are mapped to the user’s location and problematic cell is identified, basis the performance across certain KPIs. X-Net pieces together this information by sourcing KPIs from several data sources. A cell is measured on KPIs such as network coverage, signal throughput, network downtime, call health, and customer location. Frequency of usage and KPI health helps us identify the problematic cell which is relayed to resolution teams.
Impact
X-Net has provided a significant push to our complaint handling operations. We started resolving up to 4x more complaints within a 30-day period, as compared to earlier. We even reduced the reopen rate by ~50% which validates the increased efficiency in the resolution process brought about by X-Net.
Way Forward
While it has set a solid foundation, certain challenges persist in our complaint handling system. More specifically, as X-Net currently analyses top cells, it fails to identify problems in cells which are used less frequently or in cases in which user location, changes often. To solve for such problems, we are now attempting to layer X-Net with a machine learning capability. X-Net will dynamically tune itself to more parameters and classify complaints with higher accuracy beyond rule-based algorithms. The objective is to further reduce human intervention, improve the system and ultimately reduce time of resolution for complaints.
X-Net is a prime example of how we have internalised data and are now using it for the customer’s benefit. Data & Analytics facilitates decision making for different functions from Marketing and Product to Network Operations. We leverage data for use-cases from designing sharper campaigns, refining digital properties, to identifying operational inefficiencies. We, at Airtel, in the past 2 decades, have continuously innovated and has made its presence felt as a standalone tech. ecosystem with product offerings from mobility to music streaming. X-Net is only one of many examples of innovations. Next-gen CLM (article link) is another such example also driving improved experience for our customers.
