Offering conversational platforms that can think, reason and communicate like humans, artificial intelligence (AI) start-up KAMI is helping businesses automate customer service with its online virtual assistants. Through ongoing conversational analysis – it analyses and remembers user’s personality, preferences and conversational context – the machine helps collect customer intelligence and encourage better corporate strategies and product planning.

Its products, KAMI Relate and KAMI Analytics, run across major messaging platforms, including Facebook, Skype and WeChat and support multiple languages. “They enable virtual agents and chatbots to understand, empathise and build insights from conversations, thereby assisting companies and brands to grow market reach, build customer loyalty, maximise efficiencies and make informed decisions,” according to CEO Alex Cheung, who co-founded the company with former colleague, Simon Ho in 2015. The two Hong Kong natives met while working for the mobile marketing firm cherrypicks, where Mr Cheung designed the world’s first mobile instant messaging service in 2006.

“It’s a perfect tool for saving cost from customer service and assisting companies to identify potential customers and generate leads to improve return on investment,” he adds, saying that an intuitive interface provides personalised feedback to keep customers engaged. This ability to hold meaningful conversations can help customers make better decisions through deep conversational analytics.

Machine Reasoning

At the SmartPlanetNI’s start-up demo day last March in Belfast, KAMI earned the title of most investible company.

To stand out from the competition, Mr Cheung believed that focusing on data or algorithms wasn’t enough, so he set about developing machine-reasoning. Targeting the banking and finance sectors with its patented technologies, the CEO describes them as the [community-based traffic and navigation app] Waze of AI. “When Waze combines GPS data with maps, it generates traffic condition information for route and city-planning. By combining conversation dialogue with customer context and decisions, KAMI captures human decision-making process and makes human intelligence truly transferrable,” he explains.

With its regional headquarters in Hong Kong, the company is currently based in the United Kingdom after being admitted to the Startplanet accelerator programme there earlier this year. KAMI has also attracted plenty of attention in the Greater China region and beyond, including at the SmartPlanetNI’s start-up demo day last March in Belfast, where it earned the title of most investible company. It has also been shortlisted in four categories in the AIonics Awards in London, and won the Best Startup award during ARM demo day in Shanghai in June. So far, it has found investment from C-level executives from the AI, robotics and banking industries.

Common Law Advantage

KAMI is educating various industries about AI by joining conferences and speaking at events to help future clients better understand the technology and services offered.

With offices in Hong Kong, Belfast and London – the Hong Kong office is focused on core technology and software development while London is devoted to R&D – Mr Cheung says its main challenge comes from recruiting talent specialising in this particular area of AI. “As part of the UK Global Entrepreneur Programme, we’ve received support in connecting with students and graduates in the UK. At the same time, we have connected with ARM, the world’s largest processor manufacturer, in Shanghai, which will help us set up our research lab in the city,” he says.

As for the benefits of setting up an AI business in Hong Kong, Mr Cheung says, “Hong Kong is the only city in the Greater China region to implement common law, and this gives business entities unprecedented advantages in doing business in Hong Kong. Secondly, though the cost of living is high, the low tax rate and simple tax system makes it a perfect place for start-ups.” Being located in Hong Kong means KAMI can act as a technology hub within the Greater China region, providing services to companies in Asia-Pacific, he adds.

AI Education

People tend to associate AI or machine learning with big data, Mr Cheung says, and that many generally believe AI is a game for major technology companies like Google, Amazon, Baidu and Alibaba, which all own large amounts of data.

But he says that AI has implications for all businesses. “To borrow a quote from [the late US scientist] Roy Amara: people tend to overestimate the effect of a technology in the short run and underestimate the effect in the long run – and so do investors. So in the past few years, market adoption of these technologies has been really slow,” he says.

KAMI is educating various industries about AI by joining conferences and speaking at events to help potential clients better understand the technology and services offered “We provide a consultancy service that helps them study how they can make use of their own asset [and leverage their] data to solve their business challenges.”

Bright Future

Alongside blockchain and Fintech, Mr Cheung believes AI is among the hot topics in the IT industry, and expects to see more start-ups emerging in this area. He also anticipates major acquisitions within the market in the coming months. Google, Intel and Apple have been competing to acquire private AI companies alongside Uber, Samsung and Ford, the latter of which recently acquired Argo AI, a start-up that has created a self-driving car, for US$1 billion.

Amid competition among multinationals to acquire AI startups, Mr Cheung also expects to see AI companies reach a greater level of maturity. “Start-ups that have established themselves as experts in a particular vertical with customer support will start refining their offerings with the data collected, and we can expect to see technological advancements like one-shot learning [where something is learned from a single example], zero-shot learning [the ability to solve a learning task without receiving previous examples] or self-evolving systems [robots manufacturing themselves through generative design],” he says.

Over the next 12 months, KAMI will focus on working with clients in the Asia-Pacific region to enhance its offerings based on customer feedback. A full-scale launch, meanwhile, will follow in the second quarter of 2018.