Language technology company, KeyPoint Technologies uses AI, Natural Language Processing (NLP) and machine learning to predict what the users want to type on a QWERTY keyboard even before they type. The technology is available in 140+ languages, including 28 Indic languages, and it is used by 15 million users. Going forward, KeyPoint plans to capture the speech solutions market, design chatbots for local Indian languages and foray into global markets.
Sanjay Patel and Sunil Motaparti founded KeyPoint Technologies in 2004 as a company that develops simple word prediction engine to help users type faster and get sentences faster into their messages. Back then, the technology was introduced for physical keyboards. However, as smartphone penetration picked up, the company researched on language technology and eventually launched page prediction, sentence prediction and Emoji prediction services. “When we started making headway into the smartphones segment, we brought in the entire technology for it,” adds Nutan ChokkaReddy, KeyPoint’s CEO.
Soon after its foray into smartphones, the team felt the urge to do something different from what the other keyboard players in the market were doing. This was in 2013, when the company started looking at ‘intent’ more closely. In other words, instead of word prediction, it started doing intent prediction. “We do this on the device and we don’t share any information on the cloud,” says the CEO, explaining further that, “there are other intent engines that require a user to send information to the cloud. But we, being a keyboard player, were concerned about user privacy. So, we did it on the device without sending information to server.” This privacy component acts as a major differentiator for them.
KeyPoint tested this idea for almost a year and half and in 2015, launched its beta version with just 500 to 1,000 users. Over the next six months, it built on top of the ecosystem and signed partnerships, eventually taking its user base to two to three lakh. “Then OEMs saw value in this product and companies like LAVA, coolpad and INTEK started making our keyboard a default option,” says ChokkaReddy. Since then, its user base has grown to 15 million installed base with a monthly active usage of close to six million. “We had been focussing on intent and wanted to prove it on Indian audiences first. Now that we see the interest in a product like this, our idea is to take it global – US and China – by end of 2018,” says he. LG phones uses the company’s prediction engine in its keyboard, and now KeyPoint aims to covert the Xploree engine into a global product by the end of this year.
Most of the KeyPoint’s partners are based in India, including Zomato, BookMyShow, Amazon, CouponDunia and more. And, the company has a customised Xploree platform for the Indian audience. It has brought in a specialised algorithm for Indic languages so that word and intent predictions are more accurate. It also has a feature called transliteration, where a user can continue to use QWERTY keyboard, but suggestions will be in their native script. “Our focus for the next couple of years is to address the needs of the Indian audience by focussing primarily on nine major Indic languages,” shares ChokaReddy.
My primary goal is to ensure that our technology advances, so that it can be used by different market segments like OEMs and app developers. This apart, our main focus areas include generating higher revenues, improving our language technology capabilities, and developing good market capabilities
KeyPoint has three forms of revenue. First is the advertising revenue through its Xploree platform. Second is offering its engine as a SDK (software development kit) for other app developers to use, and third is the localisation business where the company localises (translates) the content into different languages. “We earlier used to generate a large portion of our revenues from the OEMS and developers. Only in the last six months have we started focussing on advertising revenue,” states ChokkaReddy. This year, KeyPoint hopes to generate 20 per cent of its revenue from advertisements and remaining 80 per cent from the Xploree SDK licensing (with 30 per cent within this being driven through its localisation business).
Eyeing the market
While the last couple of years for KeyPoint have been about getting users through the OEM channel, the team has now realised that even though 15 million users have installed its platform, there is not much awareness around its product. “This is something we want to focus on in this financial year,” says the CEO. On this front, the company is allocating a large portion of funds towards marketing and creating brand awareness. “We also plan to showcase our Xploree platform in international events, organise a few bloggers’ events, Hackathons and more, where we can promote our SDK capabilities,” he reveals.
Growing in numbers and more
Over the next three years, the company aims to expand globally. “The main challenge with Chinese, Japanese and Korean will be in identifying intent in these languages. A lot of effort has to be put in to get quality intent,” admits ChokkaReddy.
That apart, speech solutions is also seeing a pick up in the industry. “Not to say that text is not important, but speech is another primary interface of communication and we have built technology to address the scenario,” ChokkaReddy adds quickly.
It is also hard to understand what the user is searching for in India, because they don’t use standard search terms like users in the West. Technology has to evolve to understand this, and awareness of local language is also important to understand the customer better. This apart, chatbots are coming up in websites and banking operations, with all of them meant for the English audience. “We want to customise this for the Indian audience,” he notes and adds, “To achieve all this, the main goal for us is to get the right resources and sales team. That being said, we are also reaching out to Universities which conduct research in the language technology space to seek their help in contributing towards our research work.” With 95 employees on board, today, KeyPoint has offices across Glasgow, Cambodia, Chennai and the U.S., with its R&D team based out of Hyderabad. While the company received funding from two family offices – Motaparti and Patel family – it is now self-sufficient. However, in order to fulfil its plans of expanding into global markets, the company plans to look for further investments.
With such an elaborate mandate in place, ChokkaReddy is gearing up his team and creating a clear process to realise the organisation’s goals over the next few years.
Snapshot
KeyPoint Technologies
CEO: Nutan ChokaReddy
Year: 2004
Investors: Family offices – Motaparti and Patel
Concept In Brief
KeyPoint uses artificial intelligence (AI), natural language processing (NLP) and machine learning (ML) to predict what users want to convey even before they can type the letters of the word on the QWERTY keyboard. Its flagship product, Xploree, is a smart mobile keyboard app and a smart discovery platform. Powered by AI and NLP, the keyboard allows users to receive relevant recommendations and content while they chat, search, or browse apps. It also helps users to type in 140+ languages including 28 Indic languages. The Xploree app allows users in India to get native word prediction on the keyboard when they use English to type out words in their mother tongue. For example – Hinglish, Banglish, Tamlish, Kanglish.
