Telco grade communications platform on cloud Ozonetel acquires Yantrasoft’s speech recognition vertical in a payout model. The acquisition gets Ozonetel its team, IP, technology and customers that include a big consulting firm and an Agri-products client.
Ozonetel hasn’t raised venture capital yet. It is now expanding to international markets and is looking to raise its first round of funding for marketing in the new markets. Footprint expansion is starting with SE Asia and USA markets. The company is in talks with a few US-based investors and hopes to raise its first round in the next three to six months.
Yantrasoft offshore arm, which was founded in 2009 in Hyderabad, has been developing speech related solutions.
What is the Speech recognition engine?
Text-to-speech engine in Indian languages – Using this, business can now build dynamic IVRs in regional languages like Hindi, Telugu and Tamil
Indian dialect Speech recognition engine – This will allow automatic recognition of the spoken word. People can speak their choices and the system recognizes the words and converts to text.
Gender Identification – This allows businesses to automate targeted services and also conduct customized surveys.
Speech Biometric – Identifying a speaker based on the speech – This can be used by businesses that need security in their services, like the BFSI vertical.
Sentiment analysis – Content can now be steered based on the sentiment of the voice. This can be hugely useful in contact centers. If during a call, the sentiment of a customer is identified as being upset, the call can be directed where a manager can intervene and take over the call.
Currently Ozonetel has the offline version where sentiment is analysed after the call. The online version will be available soon.
Why the acquisition?
Ozonetel has had businesses asking for tools to manage their services better. The speech services are ready to be rolled out with a few clients. Ozonetel will start seeing revenue come in from this acquisition in the next financial quarter.
Speech recognition engine is a challenge in India due to multiple languages and local variation in languages (spoken Hindi in Hyderabad is different compared to that in UP). Training the engine to understand all the variations and support multiple languages is a big challenge.
Ozonetel will offer vocabulary recognition that works for multiple languages and dialects. The service provided by Ozonetel will offer vocabulary for multiple domains like city, state, digits. The offering will be across verticals like banking, travel, etc. This vocabulary will cover all the keywords needed for the domain. These would be tuned for performance and exposed on cloud, ready to be used in any application.
For example, a bank IVR system could greet you with, “What can I do for you today?” and the customer could just reply with something as intuitive as ‘Fund transfer’.
Speech is the future
Speaking exclusively with YourStory, CSN Murthy, co-founder, Ozonetel shared,
Speech is the way to go in a market like India where vernacular languages are many. However, the available solutions from leading players like Nuance are very expensive with a huge capex of investment even to try out ideas; forget about a serious launch. Besides, there is a lack of talent to implement the technology solution. All these challenges are being disrupted by Ozonetel’s offerings where, for the first time in India, we bring a menu of speech recognition services on the cloud. We have already launched KooKoo – India’s first Cloud Telephony platform with APIs for others to develop their own apps. We have 3000+ developers on the platform. Now they will have access to Speech APIs so that they can build powerful apps and support 10+ regional languages to address 900 million mobile users who speak mostly vernacular languages.
Ozonetel doesn’t have the first mover advantage in this speech space. Nuance and Lumenvox has support for Indian languages, but with few challenges.
Knowlarity and Ozonetel are both cloud communication players. However, Knowlarity is an IP PBX player and Ozonetel is a Cloud Communication platform allowing organizations to use it across marketing, sales and services.
Challenges in this space
‘The greater the obstacle, the more glory in overcoming it.’ ― Molière
Speech recognition engines are costly. Text to speech services cost $400 per telephone line while Speech recognition engines could cost upto $1000. Finding skilled resource to make it work is a challenge and costs a lot. Integrating legacy softwares systems is a pain, and is available with only certain platforms like Avaya, Cisco etc, posing a major hurdle for wide scale applications.
All the above limitations make the technology unaffordable till the organization has a budget of a couple thousand of dollars. This is the reason why businesses in India have not been implementing speech.
On asked about challenges, Murthy shared,
As we see it, major hurdles are scaling marketing & sales. Both call for investments. Investors in India don’t have that risk appetite. They want startup to have the scale before they step in.
The new government under the leadership of Mr. Narendra Modi is trying hard to help boost entrepreneurship in the nation. They have started a separate ministry to cater to the problems entrepreneurs face so that they can formulate policies to overcome them.
Speaking about government’s role in this industry, CSN Murthy said,
If the government can encourage innovation coming from Cloud Telephony industry, it will help the underserved SMB/SME segment to leverage and compete with big enterprises. Efficient Communication is the backbone of any business. SMBs/SMEs cannot afford huge CAPEX, neither can they manage complex technologies. Indian SMBs need innovative startups to offer Communication as a Service at affordable rates. Innovation comes when policies and regulations are eased. Government announced a billion dollar fund to support start-ups. Even soft loans will help as Cloud Telephony players need to continuously invest in R&D. Government should adopt technology from Cloud Telephony players for G2C initiatives.
Apart from proactive customer services, Ozonetel’s solutions are cloud based. This really makes it easy for integration with all platforms, IVR, mobile or desktop.
Another snag major clients face is payments. Most cloud telephony startups have very complex payment structures putting their clients in doubt. Ozonetel’s pay-as-you-go model seems interesting and helpful for startups.
While using these applications, clients have to adjust their sector specific vocabularies and employ resources for this. To overcome this and help clients save their valuable resources, Ozonetel is offering pre-tuned domain specific vocabularies.
Ozonetel made telephony application platform available to masses over cloud, now it will do the same for speech.