Introduction

The Data Annotation Report has been published by NASSCOM in February 2021 and outlines the landscape and trends in the Indian context.

AI and ML is one of the fastest growing technologies and yet we are far from realizing the true potential of it. While everyone is aware of the fact that AI requires vast amount of data to continuously learn, improvise and identify patterns, what we lack to understand is that the system requires to be fed with smart data to throw smart results. Hence, to do this, it is important to transform raw data into structured data that minimizes processing time. This is where data annotation becomes important. Data annotation, or in simple terms data labelling, is the process of labelling data available in various formats so that the machine can easily understand the input materials and come up with accurate outputs. Data labelling represents over 25% time consumed in most AI/ML projects.

The report the data annotation industry including its challenges and opportunities. It rightly strikes a balance by giving us a global as well as an Indian perspective. The global data annotation spend on third-party is estimated to increase seven-fold by 2023 as compared to 2018 and is expected to constitute one-fourth of the total annotation spend. As against this, the Indian market is valued at USD 250 million in 2020 and is forecasted to exceed USD 7 billion by 2030 with an estimated 70,000 professionals engaged in the process. On the contrary, while the global market is served by mostly three business models, crowdsourced platforms, managed service providers and SaaS-based offerings, the Indian market is dominated by the managed services platform that accounts for 65-70% of the total revenue. Nevertheless, annotation platform as a SaaS offering is gradually gaining acceptance because of its flexibility and scalability. In terms of the data annotation services that takes into account all data types ranging from text and image to NLP and computer vision, there are two major building blocks that ensures smooth functioning of the platform: a) a trained workforce and b) a data annotation platform with an analytics dashboard.

The report also focuses on the current challenges, opportunities and stage of maturity of the Indian players and provides recommendations to scale up the industry to match the global standards. While highlighting that 75% of the Indian players are in the initial and growth phase, the report identifies various reasons such as language barriers and lack of cultural context as some of the reasons restricting Indian players to access the global market.

Relevance of the Report

The report brings forth the fact that India has the potential to become the data annotation hub of the world thereby creating employment opportunities and scaling up the process of India’s AI readiness. However, given the fact that India has a long way to go before attaining maturity, roadblocks such as data privacy, lack of cultural context and the increasing need for non-English language data labelling can restrict India’s potential for full growth. In addition to this, as the industry gradually matures with automation platforms becoming increasingly intelligent, the business models also need to evolve. In a very succinct manner, the report lays the roadmap for unlocking the maximum value for the Indian annotation service providers and provides recommendations across the various pillars of the industry.

Key Takeaways

• The data annotation industry has a huge prospect in India as it is still in the initial growth phase. However, its global counterpart has stepped into the advanced phase
• The sectors that are driving the industry are automotive (autonomous vehicles), healthcare, BFSI, Retail, Media & Entertainment
• India’s ability to be the global hub of data annotation service stems from its decade-old service delivery experience
• The industry needs to address various challenges such as data bottlenecks, infrastructure constraints and absence of strong data policies to leverage the existing talent