Much of what consumers experience of artificial intelligence (AI) has been focused on front-end customer use cases. But AI has the potential for so much more. This first of a series of articles on network analytics introduces how leveraging the vast amount of available network data can advance applications of AI customer care and experience, and the overall running of communications networks.
Unlocking the power of network data through analytics and artificial intelligence
Network Analytics: Part 1
By Nikos Andrikogiannopoulos and Christopher Fergus
Today, a top agenda item for telecom executives is to use artificial intelligence (AI) to improve customer experience, reduce customer care costs, improve network stability and drive closed-loop automation.
So far, most efforts focus on machine learning (ML) use cases like image/voice recognition or utility, such as:
Customer Service Chat Bots: Uses scripted answers to simple requests and/or routes service inquiries to the correct agent type
Speech and Voice Services: Voice-enabled search & content selection for TV/video to improve or replace the remote-control (e.g., Comcast’s X1 Voice Remote)
Predictive Maintenance: Detect signals from hardware to alert for likely failure
These use cases, driven by tech giants such as Facebook (chat bots) or Apple (voice recognition), are rather generic. However, telecom companies possess a unique asset that can further the application of AI: network data.
The wealth of network information is a distinguishing endowment for telcos. The vast amount of data flowing in - from modems, switches, routers, servers, active & passive probes combined with customer info, usage patterns, purchases, and customer service interactions - creates opportunities to identify patterns in the data uniquely available to the telecom industry.
These patterns can create high-value AI use cases to improve operations and customer experience.
One example is optimizing field operations. This is a high-cost center for all telecom organizations. It can cost up to $150 to dispatch a field technician to resolve tough issues – it’s easy to see how quickly these costs can add up for companies that require thousands of dispatches per day.
Additionally, the experience of scheduling repeat tech visits, waiting for cumbersome troubleshooting processes, and dealing with multiple customer service interactions for the same problem negatively impacts the customer’s journey. In locations where subscribers have options for internet service, poor customer experience can lead to service-related churn and lower customer lifetime value.
Through our work, our clients have seen opportunities to reduce field operations costs and improve customer service, including*:
Reducing Repeat Dispatches
Action:Examine network data to understand if customers are still experiencing service quality issues after a dispatch
Benefit:Reduces customer complaints and saves service costs
Providing Field Tech Troubleshooting Insights
Action:Leverage network data to provide technicians a likely root cause of service quality issues prior to dispatch
Benefit:Ensures a streamlined, customer friendly dispatch process
Eliminating Unnecessary Dispatches
Action:Associate customer support interactions with dispatch success to identify inbound requests that don’t require a dispatch
Benefit:Route issues to the appropriate organization and avoid sending excessive, costly dispatches that have no impact on customer experience
By analyzing their network data, we expect operators to enable more industry specific use cases. Network analytics will thus unlock value as service providers build the ‘brains’ behind a more software-defined networking future and push the telecom industry to further the promise of artificial intelligence.<>
*In our series of articles focused on network analytics, we delve into these and other use cases for network data and AI - Read Part 2.