Artificial Intelligence Landscape in Singapore

Artificial intelligence (AI) is intelligence demonstrated by machines.

More broadly, AI is used to describe machines that mimic cognitive functions in humans, such as learning and problem solving. Founded in 1956 as an academic discipline, AI developments have gone a long way since and have found many applications in various fields: healthcare, automotive, finance, military, etc. The global market for AI technologies is expected to reach US$35 billion by 2025. More and more business organizations are using AI technologies to identify patterns and trends in vast reams of data, allowing them to make data-driven decisions and potentially become more competitive.

There are limitations and ethics issues surrounding the AI subject but they are not in the scope of this report.

Artificial intelligence (AI) is a new, exciting area to invest in Singapore. The Singapore government has been actively building up the local AI sector through initiatives like AI Singapore and Smart Nation. According to the Digital Government Blueprint, all ministries and their related agencies have to have at least one AI project by 2023 as one of their KPIs.

SourceSage aims to provide a comprehensive view of the artificial intelligence (AI) landscape in Singapore and to identify the gaps that the industry faces which A*STAR is able to fill in.
 

Figure 1. AI impact on rates of innovation and employee productivity in 3 years from now. (data adopted from a Microsoft-IDC study)

The Singapore government has been actively building up the local AI sector as it sees the potentials that AI can possibly bring to vitalize the seemingly matured businesses like banking. Its flagship initiative, AI Singapore (AISG) was founded in 2017 as an S$150 million national programme to deepen Singapore’s AI capabilities for the future digital economy. With Singapore’s broader Smart Nation programme, Singapore became the top location for AI and robotics investment in 2018 according to a fDi Intelligence report.

Figure 2. Notable examples of AI startups in Singapore categorized by industry
(by Victor Baffet)

As shown above, there are 3 focus sectors — finance, healthcare, and marketing — that attract the most numbers of investors and entrepreneurs to develop AI technologies around them. The findings are not surprising given the intense technicalities needed in these fields. Leveraging AI technologies to overcome their respective challenges seems to be a right call.

Figure 3. AI readiness in Asia with a scale from 0.00 to 4.00
(data adopted from a Microsoft-IDC study)

In a broader sense, Asia Pacific region is not AI-ready with inferior ratings in critical areas like Capability and Infrastructure.

The study revealed the three top challenges that Asian businesses face when implementing AI:

• a lack of thought leadership and leadership commitment to invest in AI; 
• a lack of tools and infrastructure to develop actionable insights; 
• and a lack of skills, resources and continuous learning programs.

Therefore, curating a correct organizational culture is critical to the success of AI implementation. Business leaders need to be more risk-taking, proactive in encouraging innovation and cross-functional partnerships among teams.

United States and China dominate the world of AI research. To understand the current standing of Singapore among other countries in the AI development, the team has compiled several key facts regarding the AI developments in various countries.

To keep the report findings as accurate as possible, SourceSage has gathered on-the-ground industry insights by interviewing key executives of established MNCs as well as startup companies in Singapore.

The key findings are as follows:

• Singapore is a good place to gain access to talents
• Singapore’s AI capability as a whole is debatable
• Organizational culture is a common hurdle for companies to implement AI 
• Partnerships are vital for startup companies in Singapore
• Additional funding is welcomed but not in urgent needs
• Most companies use off-the-shelf AI solutions

According to the Deloitte report in 2018, large software companies are integrating AI capabilities into the cloud-based enterprise software and bringing them to the mass market. Cloud platforms allow power AI tools and services available to a broad range of users. These tools are helping accelerate experimentation, innovation, and digital transformation of businesses. As AI is still an emerging technology in many parts of the world, its ethical implications remain unclear at times and most of the ethics issues only emerged after the implementation of AI.

Mr Christopher Yeo (Left), CEO of Sentient, and Mr Chen Huiguang
(Right), Head of Science & Research of Sentient

Sentient started out with 3 core values in place, that are; AI is meant for everyone, AI interaction with human must be natural and lastly, AI application development must be easy. Sentient has the vision to extend humanity with augmented intelligence. Its short-term plan is to help software developers to create AI applications rapidly using Sentient API which are AI algorithms. Ultimately, it aims to build a catalogue of APIs that can be consumed very fast, within 10 minutes, whereby the developers can simply execute AI Algorithms into their codes.

Their team consists of mainly research scientists, engineering/software developers and customer advisers. Even though Sentient has competitors from both AI and data sides of the business, it is a single platform that unifies both, and they have no known competitors for that.

Their clienteles are from three major sectors: urban living healthcare, wellness, business efficiency and optimization. One of the examples is HDB, who contributed datasets which Sentient could onboard to the platform and have other collaborators to make use of the data. Sentient has also developed AI CO-LAB which is a service that clients will pay for. There are 5 stages: from problem discovery and ending with a Pilot or full-scale deployment. Lastly, the platform is charged based on licensing fee

Sentient sees vertically-integrated companies as collaborators and sees collaboration as their key plan for the next 2-3 years. It wants to create an eco-system of AI algorithms creation and that requires thousands of APIs in the platforms. This system should enable the creation of many models in an automated way. For the data side, it has just launched a Blockchain Data Alliance and intended to create an exclusive membership for the Fortune 1000 companies.

In Singapore, the competition for AI talents is getting increasingly tough. Sentient has to fight for candidates from a highly limited pool of qualified people and is always looking out for AI talents. They value people who have rigorous academic training and advance degrees.

Ms Teo Peiru, CEO of KeyReply

Peiru is the CEO of KeyReply and works closely with the product team to ensure they understand the clients' requirements. KeyReply's main value proposition is to help their clients to automate processes and transactions with customers and among internal personnel. It focuses on low resource Southeast Asian (SEA) languages and help SEA companies in their digital transformation journey.

The team consists of 22 people with half of them are in the Engineering team and all of the engineers are currently based in Singapore to reduce friction in communications. KeyReply opens the idea of expanding to other areas with good pools of talent as there is a need to scale up AI talents in the region. Even though manpower in certain countries is much cheaper but KeyReply wants to solve problems that other people can't just do that easily so it helps to prevent intense competition.

There are typically 4 layers for the services that KeyReply provides:

• Fundamental layer: The AI layer which are the models that KeyReply trains for different use cases like predictions or classifications
• Second layer: The Business Logic layer which classifies the problem statements, approaches to the problems, and 
• Third layer: The Data layer whereby KeyReply uses data to train the models.
• Fourth layer: The Professional Services for project delivery

Most companies are still in the infancy stage to implement AI such that they often underestimate or overestimate the capability of AI and it is rare to see customers who really understand AI, hence customer education is very essential. The charges of the services are highly dependent on the scope of the project.

KeyReply is open to the idea of collaborating but it depends on what data, talents and resources each party can bring to the table.

Mr Djoann Fal, CEO of GetLinks

GetLinks is a platform and ecosystem that connects tech talents with opportunities across Asia. It supports people to build their skills, their connections, their teams, and their careers, with focus on tech recruitment and speed. It started 4 years ago and raised funds via 3 funding rounds with aggregated amount of USD 5 Million with JobsDB and Alibaba being part of their investors. GetLinks has opened offices in Vietnam, Singapore and Hong Kong with a total of 100 employees now.

Traditional job portals usually do not have complete profiles of the candidates but GetLinks gather data throughout the recruitment cycle. This allows real-time assessment of salary differentials in different geographical markets and provides transparency in how much a candidate should be paid and how much a company should pay.

In GetLinks, a ''skill'' can surge in one location. This implies a strong demand for this skill in this market and therefore, more candidates with the particular skill - the supply - will be known to the employers. Each candidate is assigned with a score on the platform based on their response time. This influences his/her appropriate pay and his/her visibility on the platform. It is made available using Alibaba AI technology. Currently, over 1000 placements were done. Candidates on GetLinks are much more likely and almost 7 times faster to respond as they are already on a platform with their complete profiles.

GetLinks's short term plan is to work with SMEs and corporate in Singapore, with focus on markets in Hong Kong, Singapore, Thailand and Vietnam. They target to get 2-3 million users so more data will be made available. They are also looking to hire more data scientists to get a stronger data team and are inclining to create AI talents through upskilling.