Every day, enterprise call centers have hundreds of customer service agents handling inquiries from thousands of customers. These interactions offer a valuable opportunity to gather insightful customer information that can enhance products and drive improved services. Customers share information not just about their product interests but also about life events, service issues and even their emotional mindset. For example, a bank customer asking a service representative about a fee on her statement might casually mention that her child is preparing to go to college — an indication that the family may soon need a home equity loan, a student loan or a student checking account.
Such insight can help companies identify opportunities early in a customer’s decision cycle, making it possible to cross-sell additional products and avoid negative experiences that might cost them a customer. Today, with consumers enjoying 24x7 ability to research and transact, opportunities to meet customer needs come and go in a flash.
Yet many organizations lack the capability to capture and analyze call-center interactions. We believe that it’s time for forward-looking businesses to address this gap.
Interaction analytics: Ready for the enterprise
In the past few years, advances in natural language processing (NLP) have enabled enterprise-grade solutions that offer interaction analytics. These solutions can elevate contact centers to a new level in generating customer insights. Moreover, they improve customer experience, contact center operations and quality.
Interaction analytics platforms ingest call audio recordings, then use NLP and linguistic technologies to translate audio to text and identify topics or concepts. Interaction analytics can also capture text-based communications from email, chat and social media. The technology helps organizations automatically analyze all interactions, instead of manually listening to and analyzing only 1% to 2% — typical for many call centers.
These platforms leverage artificial intelligence (AI) to generate highly accurate and relevant insight across many languages and dialects, including scanning all interactions to derive meaning from “non-talk” and “cross-talk” with unprecedented accuracy. For example, the AI capabilities in a speech analytics platform can identify topics that most often precede periods of extended silence, uncovering potential issues related to agent training or the knowledge base. Rectifying these issues can reduce handle time and result in cost efficiencies and improved customer experience.
Interaction platforms also can detect customer sentiment, and how that sentiment changes throughout a call, by evaluating word choice, rate of speech, tone and context. By associating sentiment measures with topics and agents, companies can develop and track improvement strategies.
Successfully implementing interaction analytics
We’ve worked with large organizations to implement interaction analytics, and the results are impressive. For example, one leading fast food chain wanted to transform its call center by focusing on four key challenges:
- Improve the customer experience.
- Reduce repeat calls.
- Understand key drivers of calls.
- Reduce call handling time.
We helped the company deploy an interaction analytics solution that analyzes 100% of all contact center interactions, including voice, email and text. As a result, customer satisfaction improved from 92% to 99%.
To achieve this, we leveraged interaction analytics to identify call drivers (reasons for the call) that had high negative sentiment and identify approaches to address the underlying need for the call. In this case, calls to resolve malfunctioning point of sale (POS) devices were resulting in high negative sentiment. By identifying a need for procedures to prevent POS malfunctions in store locations, we helped alleviate the need for the calls and reduced negative sentiment for this driver by 50%.
Non-talk time (long pauses on a call during which an agent may be researching a resolution) was reduced by 10%, which also shortened call handling time.