Case study:

Automated Technical Support Provided to Cable TV Subscribers of a Major Cable Provider

Challenge

Providing effective technical support for cable television subscribers is not an easy task. One caller’s description of what they are experiencing may be quite different from another caller’s understanding of exactly the same issue. Visual perception converted to a verbal explanation leaves plenty of room for interpretation by call center agents or any type of automated solution.

This large MSO’s primary goal was improve the response time for video technical support calls. A lack of available, well-trained call center agents was challenging customer satisfaction metrics due to longer hold times. The caller experience was also a concern with reference to how callers would interact with technical support. This cable provider was not interested in simply voice-enabling FAQs, or presenting a message instructing every caller to restart their set-top box and “call us back” if you continue to have a problem. Rather, an interactive experience was needed with natural language to assist with pinpointing issues that might be ambiguously described.

Factoring strongly into the investment analysis was cost savings. The cost of an automated technical support solution had to be considerably less than expanding contact center agent count and agent systems build-out.

Solution

SpeechCycle’s LevelOne Video Agent was selected due to its proven ability to drive to the root cause of a video problem based on caller descriptions that can range from completely vague such as “bad picture”, to somewhat specific such as “I see an error code on my set-top box”. The LevelOne Video Agent was deployed for the first call center at this MSO in approximately 90 days.

LevelOne Video Agent utilizes natural language voice recognition to capture the reason for a call. Callers initially hear a prompt that asks them to “please describe your problem in one short sentence”. A natural language model created by SpeechCycle from years of experience and collection of hundreds of thousands of utterances is able to identify the most likely call resolution path.

For example, “I have no picture” will trigger a dialog to determine whether the caller sees blue, black, or a snowy screen. Each points to a different trouble shooting path. Another example, “It says Explorer Set Top Not Authorized”, immediately tells the LevelOne Video Agent that the caller is referring to an error message on the TV screen from a Scientific Atlanta set-top box, with a precise resolution path.

Over 70 video symptoms are recognized. Troubleshooting involving set-top boxes is supported from an ever growing manufacturer and model list. Subscribers are guided by the Automated Agent to physical characteristics such as set top box light patterns and error codes. Corrective action may include instructions to press specific buttons on a remote, tighten a cable connection, change VCR channel settings or turn off a VCR, add batteries to a remote, or a host of other suggestions.

SpeechCycle’s Continuous Improvement process notably came into play at this MSO when it became apparent that callers encountering a legacy DTMF menu prior to engaging with the LevelOne Video agent were frequently pressing an incorrect button. SpeechCycle quickly updated the natural language model to recognize that these “misguided” subscribers were actually calling about cable modem issues. A seamless transfer was put in place to the LevelOne Broadband Agent which was already deployed. Callers were thus unaware that they were ever misrouted!

LevelOne Video Agent is currently fielding hundreds of thousands of calls per month at this MSO, solving customer issues that include picture quality, missing channels and DVR settings. Subscribers remain engaged with SpeechCycle’s Automated Agents for a remarkable average of 7 minutes due to the specific questions and responses that are perceived by callers as personalized for their problem.

Callers are always in control, with an ability to have instructions repeated or reach a call center agent at any time. For calls transferred to live agents, integration with CTI provides detail to a screen pop of what has already occurred as well as a next-step suggestion to save the agent the time needed to scan the list of diagnostics already tried.

Results

Benefits to the MSO continue to be: