Most people today—whether they are your customers or employees—want intuitive, frictionless experiences when they use your voice-driven solutions.
Automatic Speech Recognition (ASR) can help you save time, ramp up efficiency and scale services in your contact center and many other scenarios. However, not all ASR solutions can deliver experiences that measure up to customer expectations. Those that are too slow, too limited or riddled with inaccuracies can negatively impact purchasing decisions, erode customer trust, and discredit your brand.
It is therefore critical that you get speech automation right. And if you don’t, someone else will: Experts predict that voice commerce will grow to become an $80 billion industry by 2023.
That said, there are many hurdles in this advancing technology field. One of the key complexities is managing and expanding different dialects within a single language model. Even when two users speak the same first language, their dialects (US English vs. UK English, for example) must be managed as two completely separate languages by conventional ASR engines. These models must rely on dialect-specific lexicons and phonetic training, which can be resource-intensive, and quite expensive, especially if you serve a diverse and/or global audience.
To address this and numerous other issues, LumenVox has upgraded our ASR capabilities to bring you a new ASR engine with Transcription that is more accurate, efficient, and scalable—so you can maximize your return on investment.
Build world-class customer experiences with the LumenVox Automated Speech Recognition engine
At LumenVox, we have spent over two decades advancing our speech recognition capabilities through the development of our ASR engine. With state-of-the art transcription capabilities that are built on the foundations of Artificial Intelligence and Deep Neural Networks (DNN), LumenVox’s newly updated ASR engine offers several benefits that will help you offer seamless, accurate, super-scalable voice-enabled solutions.
With LumenVox, you can create user experiences that raise the bar—via your IVR (Interactive Voice Response), chatbot or virtual assistant.
- Improved accuracy
Our new ASR engine offers a low word error rate (compared with competitors), for exceptional reliability and accuracy. That means you can build high-performance applications time and time again.
- Incredible scalability
While other ASR engines on the market treat each dialect as a separate language, LumenVox supports multiple dialects with one language model. Take English as an example: where conventional solutions recognize US English and UK English as two separate languages, our ASR engine supports multiple dialects (including US, UK, Australian, South African and Indian) within one “universal” English language model. The same approach is true for Spanish (Castilian vs. Latin American), French (European vs. Canadian vs. Haitian), Portuguese (European vs. Brazilian), and other languages.
The ability to understand various dialects with one model means you can serve a diverse base of customers with a high level of accuracy, faster and for less cost than alternative ASR products. LumenVox uses an end-to-end DNN transcription process when building its language models so there is no need for a phonetic lexicon to understand various dialects for a given language. This speeds up the process of adding new models and when updating existing models with new language data. Your customers can also easily extend and adapt the default language models for their own needs, without having to deal with phonetic spellings and lexicons.
- Greater efficiency
With a high compression ratio of words and phrases known as ‘n-grams’, our ASR engine allows a greater number of word/phrase combinations to be simultaneously matched to the speaker conversation. This means that an accurate match can be identified in real time. As a result, users will experience little or no latency when interacting with your applications.
- More flexibility
With a selection of licensing and entitlement options to choose from, it is easy to deploy our ASR engine for any monetization model. Options include monthly subscription and usage based. This offers a refreshing alternative to competitor licensing models, which “lock” licensing entitlement to a single monetization framework.
- Better usability
Designed for simple implementation and management, LumenVox does not require costly professional services for delivery, installation, upgrades, or maintenance. Additionally, it fits in any network architecture, including premise-based installations, multi-cloud installations and hybrid prem/cloud installations.
- Simplify the migration of legacy applications
LumenVox provides a streamlined way for customers to migrate off their incumbent ASR by simplifying the migration of grammars and confidence values to the new LumenVox ASR engine.
All these features work together to help you automate your customer and user experience in a more cost-efficient, reliable, and rewarding way.
Join Joe Hagan, chief product officer at LumenVox on August 24 at 11:00 a.m. PT / 2:00 p.m. ET as he discusses what is required to deliver meaningful employee and customer experiences through voice channels. Register now for the webinar.