I saw an article online which claims that India will be an AI superpower soon. It is true in some areas. At the same time, it is important to clearly understand different segments of AI and the current status of India. This segmentation helps in actionable ideas.

First, let us define different segments of AI:    

I. AI Vertical –

The products and services that can be defined as “AI industry vertical”. This signifies that customers are paying for AI itself, not for services that use AI. This could be the AI machines renting (Ex: Google Cloud TPU, NVIDIA Training infrastructure, etc.), or the AI software libraries such as TensorFlow. This also includes fundamental research in algorithms and mathematics.

II. AI as technology horizontal – 

Generic technology which can be applied to any domain. With AI, the existing products can be enhanced. Newer features and products could be created in the business you are in. One example could be AI-enabled drones or AI-based e-commerce suggestions.

III. AI Skill-based consulting & services -

The skilled people who can play various roles such as data scientist, data engineer, DevOps specialist, AI product manager, and solution architect. These skilled professionals may also have adjacent skills like domain skills. There is also massive work in all AI projects in defining data pipelines, data engineering tasks. Besides, people need to be able to work with global teams. Combination of the specialized process for AI projects. The scale of the organization also matters.

Let us look at India’s strengths vs others.

AI Vertical - India’s Position and Potential

In the AI vertical, the main business is computing infrastructure, and anyone business vertical with a massive need for AI. Google has a massive advertisement business, which requires AI. Hence it could develop data science products. NVIDIA has GPU, which again is AI infrastructure. Due to that, NVIDIA developed its software library TensorRT that runs only on their hardware. Similar is the case with other big tech leaders. Chinese company Alibaba has its cloud and it has its library. Alibaba claims 180 million non-Chinese users. On such a basis, it is possible to introduce products in the AI vertical.

India is not much present in the AI vertical. The primary need here is the supporting business i.e. IT infrastructure companies. On this front, probably building local companies in semiconductor and cloud service would be a good first step. ‘Make in India’ can help in building local skills. It could eventually lead to setting up local companies. AI compute sticks with applications like Industrial IoT could be an interesting start. In consumer or enterprise cloud space, it will be a while before anything significant happens.

AI as Technology Horizontal – India’s Position and Potential

This is about technology insertion in industry verticals. Suppose a company has an e-commerce business, AI could be used for enhanced products and services. Likewise, if a company makes sensors, then there are opportunities to create AI technology insertion.

Hence, for becoming an AI technology leader in healthcare, India needs to have healthcare technology companies. To become an AI superpower in transport, transport infrastructure companies are needed. Only on top of successful companies with known products, AI can be added.

India has good position in a few industries. But a lot more companies are required in many different verticals. Only then, India can think about becoming AI powerhouse in those verticals.

AI Skill Based Consulting and Services – India’s Position and Potential

AI is a specialized technology domain. While there is a lot of training material online, it can only give some understanding of AI. Even if someone learns machine learning concepts, libraries, and tools, there are a lot of challenges involved in applying the technology to create a product. This is because its scope involves end-to-end analysis of user actions, taking data without creating user experience hurdles apart from the core data science. Many problems require the creation of ground truth (labeled data) for the purpose of creating training data set. This could involve manual labeling of the data, creating auto-labeling tool and data collection tools.

For some products and services, the starting point may contain no AI at all. It may start with only data collection in the first release, with non-AI features. Then iteratively product releases to be planned to include AI. This is a multi-year process. It is hard to articulate revenue due to AI. Getting returns from AI on the same year of starting R&D is almost impossible. Many companies may not want to make visible big moves by creating large in-house departments. Knowing whom to hire first is a big hurdle.

This is where the companies require AI consultancy. Technology insertion of AI involves a lot of domain and solution consulting. There is a strong need for program management skills too.

India’s software consulting with established global relationships is a big advantage in this segment. Due to this reason of breadth, depth and scale, it is not easy for other countries to compete.

India-based R&D units of MNCs is another great strength. This brings the critical combination of domain knowledge and AI skills.

Another item in this segment is AI education. There are many standard AI education courses that are dominated by technology creators. However, there is a limitation due to the diversity of the audience. Depending upon the industry, there is a need for customized live training. There are many customers who seek a package of consulting and training customized to their domain. This is where India can shine.

To sum it up,

India is already in a strong position in AI-related consulting and services. But other segments of AI would require work from the ground up. The starting point is the industry vertical and domain, not necessarily AI.

Insights obtained from hands-on experience in consulting industry (5 years), AI/ML training (2 years), Product development & procurement of ML products (5 years), autonomous driving industry experience (4 years)