Forty 7 Ronin services have incorporated mainstream industry standards and leading methodologies into our practice for the speech industry
We utilize industry recognized development standards including Function Point Analysis, Unified Modeling Language and User Experience Design. We use reusable code templates based on field tested and proven designs. This is not just “Buzzword Bingo.” We have been calibrating our function point model in the speech industry for 10 years. We have successfully used UML in the speech industry, allowing us to model your entire system in one place, to include the IVR, the database and the web services and how they interact. This eliminates the need for chasing three to five different documents.
We specialize in:
Our VUI Designers are experts in analyzing the business use cases, and translating those into call flows designed to guide the caller experience in as few branches as possible. The VUI designer stays with the entire project from Requirements gathering to Implementation. After the application is completed, it is the designer’s job to perform a walkthrough of the user interface, making sure that the finished product matches the intent from a VUI design standpoint.
At Forty 7 Ronin we specialize in building custom dynamic grammars to meet the customer’s needs. The grammar determines what the caller may utter to the IVR, which will be recognized by the speech engine. Building grammars that work is a little bit art, and a little bit science. Successful speech recognition depends on having the grammar file constructed correctly, for the caller’s situation, so the IVR can understand the caller’s response.
We are experts in crafting an optimizing Statistic Language Model (SLM) and SRGS grammars for your IVR solution.
Do you know what people are saying when they call your IVR? Forty 7 Ronin offers one of the fastest and most accurate Transcription services available. Listening to the call is important for determining what is happening in a given call. Additionally, those call listening activities can be paired with log analysis to render data driven enhancements to system performance and user experience. There are a number of services that can be performed at the time of transcription including:
Transcription – Listening to an audio file and writing verbatim what was said.
Demographics – Capturing information about the caller (e.g., gender, native speaker, accents, pitch, ambient noise).
False Rate – Comparing the transcription with the speech recognition log to determine if the utterance was falsely accepted, falsely rejected, or misrecognized.
In/Out of Grammar – Comparing the transcription with the grammars for that dialog state to determine in grammar and out of grammar utterances, then setting a flag accordingly.
Think of IVR tuning like this: Why buy a new car when an oil change and tune-up will make your old car roar?
IVR tuning is critical to successful IVR performance and customer satisfaction. We will ensure you’re at the top of your game. Speech recognition projects are complex, especially if they are done right. However, like your car, if you get the right mechanic to spend time on it, small changes can make a big difference and keep you screaming down the road of success.
In an economy like ours, resources are more precious than ever. Regular and Ad Hoc tuning present companies of all sizes the opportunity to move the performance needle in the right direction without needing to completely redesign an application or dialog state. The Forty 7 Ronin have worked on some of the largest natural language speech recognition projects in North America.