By Carmit DiAndrea, Analytics Consultant
Speech Analytics is not the thing that should be keeping Quality Assurance & Monitoring professionals up at night. It shouldn’t even be among the top five things that keep them up at night. While Speech Analytics applications provide unprecedented access to previously untapped intelligence in calls, the way these technologies work and their existing limitations make Speech Analytics applications far more valuable when pointed towards other use cases – customer satisfaction, customer and agent retention, call center efficiency, call driver understanding, repeat call patterns, call deflection opportunities, assessment of digital channel performance, identification of processes in need of re-engineering, and more.
Speech Analytic applications have robust capabilities that leverage AI and Machine Learning to understand the relationship between words uttered and uncover conversational topics, to understand human emotion (sentiment) being expressed and even leverage data about the customer and predictive analytics to identify likely outcomes. However, Speech Analytics applications on the market today do not have the sophisticated logic trees that would be required to follow and assess whether an agent conducted the correct troubleshooting steps or due diligence to uncover sales needs, is challenged by the nuances and cultural variations in human interactions and is ultimately limited to the spoken word. Imagine a sales queue in which an agent needs to ask a series of questions to uncover customer needs, reference the answers to those questions against a regional pricing matrix in a knowledgebase article, and also take into consideration a competitor’s promotion or monthly budget communicated by the customer earlier in the conversation. A human QA professional can navigate that logic tree with relative ease and assess whether the offer made by the agent was appropriate, but today’s Speech Analytics solutions would fail entirely in assessing the appropriateness of the offer made. Current speech analytics solutions are also challenged in accurately interpreting sarcasm, in understanding cultural differences in communication styles (for example, “yelling” as a normal mode of communication) and are heavily impacted by audio quality and background noise.
So instead of fearing the impact Speech Analytics might have on your QA practice or team, start thinking about how you can leverage the technology to amplify the value your team delivers to your organization:
- Evolve past assessing a random selection of calls. Random call selection has been a foundational practice in quality assurance since the inception of the practice. The roots of this practice are grounded in statistical population sampling and ensure that the results generated through QA are representative of the entire universe of calls handled. But we all know that not all calls nor customers are created equal. What if you could focus your QA efforts on high-value customers? Or high-risk customers? What if you could focus your QA efforts on a particular supplier or piece of equipment that is failing in the marketplace? Or to help the business quantify the impact of a change in a business rule? Speech Analytics can help you be more targeted in your QA efforts.
- Automate where you can. The fact is that Speech Analytics performs incredibly well when tasked with assessing objective “black and white” attributes that can be summarized into simple if/then statements. For example: If call type = ABC, then agent should use branded call opening 1. If agent sold digital phone service, then agent should mention TPV or third-party verification. Compliance-related questions tend to fall into this category of “black and white” objective attributes. By leveraging Speech Analytics to assess these attributes, these behaviors can be assessed on 100% of calls (not just a random sample of calls assessed by a QA team member) and that QA time can be re-allocated to higher-value activities (more targeted listening, as recommended in item #1 above, coaching, training, mentoring, etc.)
- Turn your QA team into a Business Insights “feeder” group. Typical QA practices focus on agent performance management against a defined set of behaviors or attributes. But customer conversations do not follow a script. During any conversation, a customer might be complaining about or praising your organization’s self-serve portal, your IVR or chatbot, a product or service, pricing or even a marketing campaign or newspaper article about your company. All of that information is intelligence that is critical to key stakeholders within your organization. And while Speech Analytics will be needed to quantify such feedback at scale, the Speech Analytics team needs to know what to search for – conversational topics, key words and phrases, call types, etc.
The rapid acceleration and proliferation of technology innovation can be overwhelming at times, and change is uncomfortable. But when it comes to Speech Analytics and Quality Assurance, complete replacement of humans by machines is at least a decade away. In the meantime, there are many Speech Analytics capabilities that can actually benefit and increase the value of your Quality Assurance practice, one of which may be an alternative career path for your QA team who may have chosen quality assurance because of their love of the energy, productive hum and “hear it first here” characteristics of the contact center floor.
Carmit DiAndrea is a contact center, analytics, and Speech Analytics expert with over 20 years of experience translating customer feedback, operational, and contact center data into business strategies that transform organizations. Carmit is also an educator, teaching statistics, research methods and algebra as an adjunct instructor at Kaplan University and is the lead instructor of the Speech Analytics MasterClass (speechanalyticsmasterclass.com)
Concentra Solutions maximizes contact center performance through Quality Assurance, specializing in Compliance and Customer Experience. We partner with internal and outsourced contact centers to provide insights and recommendations to drive performance and operational improvements. For more information, visit www.concentrasolutions.com.
