Gyst Product Overview

Gyst uses machine learning and advanced algorithms to measure, analyze, and personalize the voice experience in real time. You can skip right to the technical implementation details here or watch the short overview explainer movie below.

Gyst Technical Overview

Gyst adapts your voice channel to suit the demonstrated abilities of each individual customer as they navigate their way through your dialogue flows. The voice channel here can be your voicebot, your interactive voice response (IVR) system, or any means you use to communicate with your customers using voice.

This process of measuring, analyzing, and personalizing is accomplished via the Gyst recommendation engine as illustrated below.

Gyst Recommendation Engine

The Gyst recommendation engine is server-based code accessible via the Gyst web API. This software has been developed and production tested on hundreds of millions of voice sessions to date.

When run for the first time on a particular voice application, Gyst gathers behavioral information on how your users currently use and navigate their way through the application flow. Thousands of samples are taken at each unique interaction point to determine how well (or poorly) your customers respond to requests for speech and/or DTMF directed dialogue input, natural language interaction, and any response request needed to allow them to self-serve in the voice channel.

Gyst then adjusts the speaking rate (words per minute spoken), response time allowed, best mode of input (speech or touch-tone), inflection, message content and other parameters to suit each individual user as they progress through the dialogue flow. This provides a significantly better call experience and allows a greater number of users to self-serve in the voice channel.

Basis of Operation
User skills at navigating your dialogue flow inherently accounts for differences in cognitive abilities, environment and background noise, mobile phone signal degradation, and knowledge of the dialogue flow itself. Every user gets their own unique experience, custom tailored on the fly based on their demonstrated abilities at a given interaction point. Users that are skilled and accurate at navigating the voice application hear the voice prompts at gradually, almost imperceptible, increasing playback speeds. Slower, less skilled or less knowledgeable users (distracted, mobile users in poor signal areas, elderly) hear the dialogue progress at a slower pace. Gyst accommodates and tailors the experience for everyone in between too.

This adjustment is not a one-time or static occurrence. Skill reassessment and re-adjustment takes place every time the user responds to a voice prompt. In addition, with Gyst's advanced features, struggling users can get additional instructions and additional time to respond, while faster users may hear abbreviated instructions in order to further speed them through the dialogue. This provides the best opportunity for your customers to self-serve in your voice channel – and to remember to use it again the next time they are given the option to.

Product Features

Below is a brief summary of Gyst's features and the benefits they provide.

Anonymous - Secure
Because this personalization is driven via the users exhibited behavior, and not using stored, static user profiles, there's no need to share account numbers, PIN's or any other type of personal information to achieve significant CX and self-service or automation improvements. This provides a level of personalization above and beyond more traditional, profile based CX improvement techniques.

Adaptive Playback Control
Gyst gradually speeds up audio playback (wpm) for skilled/expert users every time they respond well, and slows it down for struggling/novice users every time they respond poorly, relative to other users navigating the same voice application.

Personalized Response Times
This allows extra time for struggling users to respond to your voice prompts. It helps reduce those annoying (and time consuming) "I'm sorry, I did not understand..." messages. Your customers will appreciate that and be more likely to use the voice channel the next time they are given the option to.

Skill Based Prompting
This feature provides additional audio instructions for users struggling in the voice channel. Sometimes just a few additional instructions or a more empathetic sounding voice can help your customers to self-serve.

Application Flow Optimizer
This feature proactively slows down audio playback at any and all interaction points known to cause difficulty for the majority of users. This has the effect of smoothing out the call flow for all users and helps reduce overall input errors.

Advanced Interaction Analytics
This feature exposes customer pain and friction points in the dialogue flow now, and as your voice applications change in the future. Our machine learning and advanced algorithms factor in user skills, response times, aggregate error rates and more – all in real time. So there's no need for you to wait for your analysts to study your static, traditional IVR reports anymore.

Intelligent Opt Out
This feature tracks each individual users performance at navigating the dialogue flow and gets them to an agent before they get overly frustrated in the voice channel. Alternatively, this feature can be used to trigger a chat or email session with the user if your call center technology supports this.

Smart Input Switching
This feature allows your voice channel to offer speech only, touch-tone only, or both based on what your users demonstrate works best for them at each unique interaction point in the voice application. Your customers will always get the input modality that works best for each situation.

Implementation Overview

We have designed Gyst from the ground up to be secure, reliable, and highly scalable. We also designed it to be extremely fast and easy to use in new and existing voice applications - both premise and cloud based. Any voice application can be optimized with Gyst by simply adding http calls to the Gyst API via the voice application itself.

Gyst has fault tolerance built in. Though the service is highly reliable and scalable, if it were somehow to become unavailable, your voice applications would simply continue to run without personalization, just as they do today.  

Gyst also provides high-resolution behavioral analytics for deep insight into how your customers use your voice channel and how and where improvements can be made. This not only provides opportunities to find and clean up old legacy application pain points, but it can also be used to guide your digital transformation strategies. Knowing what works and what needs attention in your voice channel is a great way to gain valuable insight into what works and what doesn't work for the customers you've already paid your marketing and sales teams to acquire.