The Benefits of Application Tuning

Monday, June 5, 2017 - 16:30

Nigel Quinnin

Speech has played an important role in the ecosystem of self-service options available to your customers, whether your implementation is a simple Interactive Voice Response (IVR) system or a robust omni-channel solution. Customers choose self-service when they are in search of information, making the quality of their experience paramount on all channels - including Interactive Voice Response (IVR). You may have, as a result, heard customers grumble about the automated system not understanding them, or sending them to the wrong extension - or maybe you're the one doing the complaining!

ASR often catches the brunt of the blame for these issues. Ok, sometimes the ASR gets things wrong and there can be a number of reasons for this, but you'd be surprised at just how much influence application grammars and settings have on the solution's overall performance.

One of the biggest challenges we see today is getting buy-in from application developers, system integrators and most importantly, end users to perform on-going speech application tuning. This is mainly because there is both time and effort involved in performing a tuning cycle - someone needs to collect the data, transcribe it, and analyze it; once the analysis is complete, changes then need to be effected.

As a core technology provider of speech, we've both seen and demonstrated time and again that the benefits resulting from performing even a minor tuning cycle easily and usually far outweigh the costs.

Something else we've noticed is that even applications that have benefited from an initial tuning cycle are often not treated to any follow-up tuning analysis (A.K.A. application maintenance). This creates sub-optimal performance, as it generally takes several tuning cycles and ongoing collection of incremental data (utterances) collected over time to see patterns that emerge and effectively address problems that users encounter.

LumenVox recommends performing tuning on applications somewhat regularly over the life of an application: there are almost always things that can be improved by analyzing data and applying changes based on a new data and a fresh tuning analysis. As many of our partners and users already know, the LumenVox Speech Tuner is available at a very reasonable price, and it generally pays for itself many times over - often after a single tuning cycle. But we'll leave that for another day …

In conclusion, I recommend engaging a LumenVox Skills Certified Partner to perform at least one tuning analysis for every application deployed - if for no other reason than to confirm it is working correctly. Also, consider adopting a methodology where tuning cycles become a routine part of an application's maintenance - the price of the tools is not prohibitive, and the benefits are potentially very great.

Recent Blog Entries

Jeff Hopper

As a foundational component supporting our channel partners and partner-focused business model, LumenVox offers a Partner Skills Certification Program. The Skills Certification process includes a training class and exam. The training provides basic knowledge transfer around LumenVox software and related tools. In the class, we provide partners with an overview and basic principles of how speech recognition and text-to-speech synthesis work...

Brian Pia

In a recent post; the ROI of Speech we discussed ways in which the use of speech recognition technology has changed the face of how companies interact with their customers. Perhaps the most significant benefit realized through the implementation of a speech enabled service solution is the enhanced level of intelligence delivered during the customer interaction. Customers are no longer bound to pushing keys to force fit their call reason into the company’s pre-determined options.

Maria Simonton

As mentioned in our previous blog, tuning analysis greatly enhances the performance of speech-enabled applications. LumenVox has invited Interactive Northwest Inc. (INI), a LumenVox Skills Certified Partner, to share their insights on the process of speech tuning and the improvements associated with ongoing tuning over the life of the solution.