With recent advancements in Speech Analytics technology, the Contact Center is no longer just a cost center. Instead, it has become the go-to source for your Voice of the Customer (VOC). Harnessing this VOC feedback can grow your business, as consumer loyalty becomes increasingly dependent on customer experience, and less on product and price.
Customer service makes the difference between whether your customer experience inspires loyalty or encourages churn. Ensuring that your agents are well-trained and knowledgeable is key to success — that’s why every contact center has coaching, training, and quality teams to help agents improve. Traditionally, agent coaching has been a slow, manual process. Coaches take the time to listen to 2 to 5 calls per agent, per month, to evaluate call center employees. It is a large investment of time for coaches, and only includes a small sample size on which to judge an employee. Machine learning and other big data technology can right this process and can make coaching faster, more scalable, and near real-time.
With machine learning, you can listen to and score 100% of all your interaction data, instead of only relying on a handful of phone calls. In place of just phone interactions, you’ll have the omni-channel approach with customer texts, emails, and reviews, all informing your evaluations. Software with a robust analytics engine empowers you to create automated rubrics around those channels, and analyze the aggregated behavioral patterns. The analytics engine does the tedious work of listening to all customer-client interactions and extracting meaning through deep understanding of customer intent and agent reaction. This frees up coaches to find the right coachable moments across those millions of moments, ultimately reducing the time taken to coach, improving agent performance, and increasing customer loyalty.
By deploying an omni channel analytics engine to power your Agent Coaching program, you’ll be able to:
- Analyze 100% of your interaction data across all channels
No longer will QA be dependent on a small sample of your calls, or a large team of QA supervisors. You can use interaction analytics to monitor every interaction — spoken or written.
- View speech and post-call surveys in the same view
The advantage of transcribing all your phone calls to text is that you can mash up this data against your other text-based sources. Your supervisors can look at customer feedback from the post-call survey, while listening to the actual interaction on the same screen, in the same system. This holistic scorecard provides a comprehensive view on team and individual agent performance.
- Use Natural Language Processing (NLP) to view “moments of truth” or “peak emotion” during a call before listening to it
Supervisors and coaches spend a lot of time listening to calls looking for coachable moments. Natural Language Processing and machine learning technology reduces that time and takes you right to the relevant information. Once you convert a call to text, you can overlay it with customer sentiment, effort, and emotion and pinpoint the exact moment in the call where the customer felt “peak emotion.” You can then show the snippet of this call on your dashboards to provide supervisors with context before they listen to a call. This dramatically reduces the time to find actionable interactions to coach on.
- Create weighted rubrics based on soft skills and business rules measured against a benchmark
Are there agent behaviors that result in higher customer loyalty? An analytics-based engine lets you weight those behaviors highly in evaluation rubrics, to encourage the best CX for customers. These behaviors can range from soft skill measurement (empathy, proactiveness, professionalism, etc.) to call handling skills (hold management, transfer management, escalation management, etc.) to compliance (disclosures, verifications, authorizations, etc.).
- Measure not just what was said on the call but what was not said on the call
Most QA scorecards are based on what agents did. However, it’s just as important to find out what agents didn’t do. With an analytics engine, you can run every interaction through a framework that looks for the “did’s” you want to encourage, and then flag areas where these occurrences didn’t occur. And each “didn’t,” is a coachable moment.
- Curate your own calls for listening and coaching
Supervisors and coaches often want to look for calls based on ad hoc criteria. There might be an ad hoc category not already modeled in your scorecards — say, calls under 5 minutes long where the customer was angry but the agent did not express empathy, or an emerging business event. A machine learning program can be adapted to meet these on-the-fly needs. That way, coaches can curate and tailor their coaching sessions for maximum impact.
- Apply conditional logic to your scorecards
It’s important to be able to score or analyze interactions beyond just a simple pass/fail logic. For instance, if a customer calls in about simply changing a password but is not necessarily angry or frustrated, the agent does not need to overtly show empathy. Similarly, many times, agents are only supposed to perform certain actions if the customer intent warrants it. You can model all of these if/then scenarios via a robust interaction analytics platform to truly measure the effectiveness of your agents.
- Benchmark your best agents vs agents that need more hands-on coaching
Once you’ve calibrated your rubrics to your desired outcomes, you can easily create side by side reports of your best performing agents versus agents who still have the opportunity to grow. You can compare agents on almost anything in an interaction — soft skills, call handling skills, empathy, tone, choice of words — and use this information to create “best practice” coaching manuals and agent scripts, or an entire library of exemplary calls.
- Send personalized recommendations to agents for self-coaching
Once you’re using analytics to capitalize 100% of customer interactions, you can start empowering your agents by providing regular automated feedback reports to their inboxes. These reports can contain high-level performance metrics for the past day or the past week, some examples of customer comments — good and bad — and comparisons of agents, along with personalized recommendations on how to improve. These automated reports complement in-person coaching and foster a culture of continuous improvement in an organization.
The Contact Center is your greatest resource, if you know how to use it. A machine learning solution empowers you to use Contact Center data to create a holistic picture of the Voice of the Customer. Instead of guessing what your clients want, you’ll be able to ask them, with the power of an analytics-based engine. And the more you know what your customers want, the more likely they are to be loyal.
Keegan Brenneman is the Product Manager for Case Management at Clarabridge and also works with the CX Studio Product Management team. Prior to his current role, he worked as a Customer Care Engineer for nearly three years. He may be reached at email@example.com.
Shorit Ghosh is Vice President of North America Services and manages a Clarabridge team of consulting managers, business consultants and technical architects to help his customers improve their own customer experience, increase revenue and reduce cost and churn. He may be reached at firstname.lastname@example.org.
Clarabridge helps the world’s leading brands take a data-driven, customer-focused approach to everything they do. Using AI-powered text and speech analytics, the Clarabridge experience management platform enables brands to extract actionable insights from every customer interaction in order to grow sales, ensure compliance and increase operational efficiency. For more information, please visit www.clarabridge.com/platform/contact-center/.
See Clarabridge at the Sponsor Showcase at the 2019 QATC Annual Conference.