Dear

VocalIQ was formed in March 2011 to exploit technology developed by the Spoken Dialogue Systems Group at University of Cambridge, UK. Still based in Cambridge, the company's has a B2B focus, helping other companies and developers build spoken language interfaces. Example application areas include smartphones, robots, cars, call-centres, and games.

More than a billion smart devices were shipped in 2013, with input interfaces that are difficult to use and voice interaction often available but seldom used. VocalIQ’s proprietary technology dramatically improves the performance of voice-based systems and simplifies the authoring process for new applications

The company provides a layer of middleware that sits between the speech recogniser and the application. This middleware implements machine learning algorithms which interpret and track the user’s intentions, and automatically determine the most appropriate response back to the user. 

More detail can be found on the company's website here.

Based on award-winning research from the University of Cambridge, VocalIQ uses state-of-the art techniques for all its components. These technologies have been tested in various settings, showing significant increases in performance compared to traditional approaches typically used in industry. Specific benefits include increased success rates, shorter dialogs, and reduced development costs.

• Semantic decoding: Before deciding how the system should respond, it is important to work out what the user meant by what they said. There are always many ways to express the same thing in a conversation. Deciphering this meaning is the task of the semantic decoder. VocalIQ has developed various machine learning approaches to learning the meaning of a sequence of words automatically, and it provides this technology as part of its products.
• Dialog management: Deciding how to respond to each user input is the task of the dialog manager. By integrating everything that might have been said in the dialog, including possible errors, we have been able to show significant improvements in the decision making performance.
• Language generation: System prompts and responses to questions are designed by the application developer using simple template rules. These are then conveyed to the user via a text-to-speech engine.

The Opportunity 

Based on award-winning research from the University of Cambridge, VocalIQ uses state-of-the art techniques for all its components. These technologies have been tested in various settings, showing significant increases in performance compared to traditional approaches typically used in industry. Specific benefits include increased success rates, shorter dialogs, and reduced development costs. 

Semantic decoding: Before deciding how the system should respond, it is important to work out what the user meant by what they said. There are always many ways to express the same thing in a conversation. Deciphering this meaning is the task of the semantic decoder. VocalIQ has developed various machine learning approaches to learning the meaning of a sequence of words automatically, and it provides this technology as part of its products.

Dialog management: Deciding how to respond to each user input is the task of the dialog manager. By integrating everything that might have been said in the dialog, including possible errors, we have been able to show significant improvements in the decision making performance. 

Language generation: System prompts and responses to questions are designed by the application developer using simple template rules. These are then conveyed to the user via a text-to-speech engine.

The Market

The market for voice-based interfaces is spread over several large and growing areas. Voice interfaces are becoming an important component of many smart phone applications, a market of $15.3bn in 2013 with a growth rate of 29.8%. Other markets include electronic devices for the automotive industry ($15bn), digital language learning ($1.2bn) and video gaming ($66bn).

VocalIQ’s competitive advantage comes from having unique dialogue management algorithms, significant and deep technical knowledge and sophisticated state-of-the-art proprietary software. All three founders are internationally recognised as leaders in their field.

Management / Team

Blaise Thomson, CEO, spent several years researching new approaches to building spoken dialogue systems before co-founding VocalIQ; first for a Ph.D. and then as a Research Fellow at St John's College Cambridge. Many of these new ideas are integrated into the company's technology and have been awarded prizes within the research community.

Martin Szummer, CTO, worked as a Senior Researcher at Microsoft, before joining the team at VocalIQ. His past research has included natural language processing, text mining and image recognition.  In addition to being an expert in machine learning, Martin has experience in early stage business development.

Steve Young, Chairman, is a well-respected entrepreneur and researcher in spoken language processing. He was the original developer of the HTK speech recognition toolkit, widely used in the speech community, was a co-founder of Entropic, acquired by Microsoft in 1999, and chairman at Phonetic Arts, which was acquired by Google in 2010. As well as serving as chairman at VocalIQ, Steve is currently the Pro-Vice Chancellor of the University of Cambridge and a Professor of Information Engineering.

Valuations as of 10th June2014. Net NAV includes income tax relief and fees on investment

Parkwalk’s current portfolio consists of 27 companies, which have raised in excess of £250m of funding between them since 2010 and over 600 patents protect their technology and processes.
Overall, Funds I-IV’s portfolio comprises of 9% Series ‘A’, 33% Series ‘B’, 25% Pre-IPO and 33% AIM listed.  

Past performance is no guide to future performance. The price of a fund can fall as well as rise. Investors may receive back less than originally invested.