Using Speech for Postcall Notes

Article Featured Image

No doubt, companies face significant challenges in providing an improved customer experience while simultaneously managing operating costs. To accomplish both, innovative companies will look for ways to use customer-facing automation solutions to improve performance in the back office as well. Applying speech recognition capabilities to a call center agent’s desktop is one such way to improve productivity and ultimately drive customer satisfaction.

Speech dictation of call notes during and after calls has the potential to produce a quick return in a contact center while improving the customer experience almost instantly. In a recent production trial, in-call note dictation resulted in better than a 30-second reduction in handle time per call and 97 percent accuracy.

Deploying a simple-to-use desktop application allows agents to dictate postcall notes. The training period is short, agent adoption is high, and call handle time is significantly reduced. But there are a few key points to consider when introducing speech technology into the contact center:

Does It Fit?
To determine fit, a thorough assessment of the current process for taking call notes is recommended. This assessment should evaluate the program’s operational needs, human factors considerations, and technical environment.

  • An operational review provides a high-level understanding of how the program’s business needs and challenges influence agent performance. It also provides a level of quantitative understanding around the scope and scale of the potential business impact of applying the solution.
  • A human factors review reveals the opportunities for call note automation via speech enablement. This involves a detailed review of the interactions with the various desktop tools during and after contact with the customer where notes are taken. How frequently does the agent have to enter notes and into how many systems? Are there limitations on the content of the notes?
  • Finally, and most critically, technical validation confirms the technical fit of a solution from the point of agent interaction via headset, through the network, and to the supporting applications. For example, specific Windows microphone settings may need to be adjusted, headset amplification and/or noise reduction may be required, and file size and network utilization must be evaluated.

Is There a Unique Call Center Language?
To achieve high speech recognition accuracy, it is important to understand a call center’s lexicon. A review of the call notes, specifically by call types, reveals relevant vocabulary. By establishing a baseline dictionary, higher initial accuracy and adoption can be achieved. In a recent pilot, agent suggestions that were collected via focus groups and online surveys resulted in a 67 percent expansion of the vocabulary. Ninety-two percent of agents agreed or strongly agreed with the statement that the solution would make their jobs easier.

Is There Buy-In from Agents and Management?
To optimize buy-in and show impact, a pilot team should be composed of agents already working together in the center. This engenders cooperation, competition, and support from management. Coaches, usually team leaders, are also trained on the solution and understand the best practices in using the solution from the very first day of the trial. Managers should post results and recognize top performers daily. Key metrics to incent include highest accuracy, most consistent usage on calls, most innovative suggestions, and best average handle time improvement. Finally, be sure to include an agent acknowledgment celebration at the end of the trial period and prior to expansion across centers.

Speech-enabling the note-taking process drives bottom-line impact and provides companies with another way to further their competitive advantage in an age where the customer’s contact experience continues to be commoditized. Companies can apply the time saved through improved agent efficiency to one or more business objectives. Therefore, they should contain costs by directly lowering average call handle time, increase sales opportunities by allowing agents to focus more time on product upsell/cross-sell, and improve customer satisfaction scores by eliminating the distraction of intracall note-taking and by retaining better call history notes for future interactions. 

Jo Ann Parris is vice president of relationship technology management at Convergys. She is responsible for marketing, business development, and solution management for Convergys’ portfolio of self-service and advanced automation solutions. She can be reached at jo.ann.parris@convergys.com.

SpeechTek Covers
for qualified subscribers
Subscribe Now Current Issue Past Issues