SSupported by cloud hosting provider DigitalOcean – Try DigitalOcean now and receive a $200 when you create a new account!

Yactraq’s Machine Learning Driven Audio Mining Puts Speech Analytics Within Reach Of Any Budget

Listen to this article

Yactraq’s machine learning driven audio mining offering makes custom vocabulary based speech analytics affordable for any organization, including SMB’s and enterprise business unit’s. Yactraq has been selected in a Gartner Cool Vendor (Smart Machines) 2016 among the first 20 companies to be showcased by IBM Watson’s Cognitive Gallery and is an IP for Defense partner of Intellectual Ventures, the world’s largest patent bank. Below is our interview with Jeh Daruvala, Chief Executive Officer at Yactraq:


Q: You’ve developed a breakthrough artificial intelligence platform for audio-mining and speech analytics; can you give us more insights into your technology?

A: Beyond basic functionality offered by legacy audio-mining systems, Yactraq’s OmniTraq audio mining product offers revolutionary machine learning driven capabilities that enable our clients to detect customer sentiment and monitor service quality in new and powerful ways, across a range of verticals. OmniTraq can be viewed as an analytics, reporting and search layer that depends on two key technology platforms. One is IBM Watson, which Yactraq uses for general speech vocabularies, for multi-lingual capability and for generating machine training data for our CoreTraq platform. The other platform is Yactraq’s own patent pending and award winning CoreTraq speech based semantic engine. CoreTraq uses a machine learning driven approach that quickly and cost effectively customizes its vocabulary to include product names and other terms that are unique to a specific client. This helps to deliver meaningful business insights with a high degree of precision.

omnitraq-graphic Recommended: Kdan Mobile Announces Its Global Breakthrough – 60 Million Downloads Worldwide

Q: What advantage does Yactraq have over its competitors?

A: Legacy phonetic speech systems are unable to keep pace with and benefit from rapid advances in machine learning. Their highly limited vocabularies do not allow higher-level semantic extraction via natural language processing. These fundamental limitations, combined with their low precision, drives the need for more capable and cost-effective speech-to-text based semantic engines.

Yactraq’s CoreTraq platform can ingest semantic and linguistic training vocabularies of hundreds of thousands of words and millions of sentences in a matter of hours or days, and do this in a seamless, almost fully automated manner. This degree of automation allows Yactraq to deliver custom vocabulary speech systems at costs that are orders of magnitude below phonetic systems. This reduces manual effort and costs, resulting in far more business value, to the point where almost any business of any size can afford and access audio mining.

yactrq_featured Recommended: Yocale – Changing The Entire Game Of Online Booking

Q: You’ve recently announced long-term partnership with Intellectual Ventures; tell us something more?

A: In the spring of 2015, Intellectual Ventures (IV) and Yactraq signed an IP for Defense (IPFD) contract thereby reaffirming our market position as a highly differentiated and significant enabler of speech analytics capabilities. Membership in the IPFD program makes IV’s 40,000 plus active patents available to defend Yactraq in the event of any IP related litigation launched against Yactraq by a larger speech technology player. IV signs a very limited number of IPFD partnerships, so Yactraq is arguably among the best defended speech startups on the planet. This provides Yactraq a very high degree freedom to operate and strongly assures clients they can be confident of Yactraq’s survivability as a speech tech startup.

Q: Could you explain the function of your CoreTraq platform?

A: CoreTraq’s machine learning capability is optimized for delivering custom vocabularies, specifically tuned for a Yactraq client, in a quick and cost effective manner. Not only does CoreTraq customize the vocabulary of its speech recognition engine, it also integrates the speech engine’s vocabulary with that of its natural language processing (NLP) module. During run time, CoreTraq’s NLP module uses its machine learning capability to automatically cluster and classify every minute of speech across large taxonomies. During the training phase of the system, CoreTraq’s NLP module uses a patent pending Yactraq process to automatically extract key linguistic features and train the language model of CoreTraq’s speech engine in an almost completely robotic manner.

couretraq Recommended: Spenco Medical – The Shape That Feels Great

Q: What are your plans for next six months?

A: Having been selected a Gartner Cool Vendor 2016 (Smart Machines) and being among the first twenty companies showcased by IBM Watson in its Cognitive Gallery, our focus over the next six months is to build strategic ties with key customers and marketing partners delivering solutions in the financial and retail space. Yactraq is also engaged in evangelizing market awareness of speech analytics since audio mining is perceived as being ‘out of reach’ for many businesses due to the high-priced legacy products offered by current vendors. Use cases we are currently delivering include call agent script adherence, predictive sales analytics and PCI compliance for redacting credit card information from call recordings. Over time, we expect to accelerate our go-to-market into other verticals and use cases like financial services, government, telecoms, healthcare and cybersecurity.

Activate Social Media: