数据标注：驱动人工智能革命的十亿美元级市场潜力

FEBRUARY2021 Foreword Artificial intelligence holds the key to an era of innovation and is increasingly becoming pervasive inour lives. Businesses across sectors are leveraging the transformative potential of AI for data-driven decision-making. However, at the core of the AI revolution lies the need for large trainingdatasets, something that most enterprises are struggling to address. This challenge worsenedamidst the COVID-19 pandemic as AI became a ‘must-have’ instead of a ‘good to have’, in theincreasingly contactless society, propelling a need for AI-powered solutions for pandemic responseand recovery. The COVID-19 war rooms further led to a reinforcement about the importance of, as well as gaps in,the data assets that are fundamental for successful AI solution development and deployment. WhileIndia marches towards its journey to realize the $500 bn. from Data and AI, the need for qualityannotated and labelled data becomes an area of prime importance. Data annotation and labellingservices play a critical role in bridging this gap by helping enterprises with quality training data fortheir AI models. Debjani GhoshPresident, NASSCOM This report attempts to highlight the huge potential that India has to become the dataannotation and labelling hub for the world. The study demystifies the data annotation landscape,market serviced by India, and the potential impact that the industry can create, both in terms of jobcreation and accelerating India’s AI readiness. We hope that you enjoy reading the report and, as always, we look forward to your feedback/suggestions. Contents Executive Summary Indian Players: Current Maturity, Challengesand Opportunities Current Annotation MaturityChallenges to Market AccessCOVID-19 Impact: Adapting to WFHGrowth OpportunitiesFuture Success Drivers Data Annotation Overview: A GlobalPerspective Growing Need for Data AnnotationGlobal Data Annotation Market and LandscapeBuilding Blocks: Trained Workforce + Smart ToolPreferred Annotation Business ModelsGlobal Service Provider Maturity Outlook and Recommendations 2030 Outlook: Economic Value Add and EmploymentFuture of Annotation Business ModelsRoadmap for Indian SMBsImpact on India: Job Growth + AI ReadinessRecommendations to StakeholdersGo-To-Market for New Entrants The India Story: Annotation Landscape andTrends India Data Annotation Market and LandscapeWorkforce Models: Dedicated + PartneredEmployment Impact and Skill DevelopmentThe India Value Proposition Appendix: Case Studies Executive Summary (1/2) The India Story:Annotation Landscapeand Trends Data AnnotationOverview: A GlobalPerspective SECTION I SECTION II •The AI revolution drives the need for large training datasets– which are fulfilled by data annotation services •Annotation market serviced by India in FY20 is ~USD 250 Mn.•~60% of the revenues from the US market •An estimated 70,000 professionals engaged through the full-time and part-time (freelance) employment models •Global data annotation spend on third-party solutions is•Estimated to be 7X by 2023 compared to 2018•Expected to constitute 1/4th of total annotation spend •Indian Annotation revenues are derived from multiple businessmodels •The building blocks of Data annotation services cover•Trained workforce•Efficient tool to operationalize labelling •Managed Services ~65-70%•Crowdsourced Platforms ~25-30%•SaaS offerings ~5% •Global data annotation landscape comprises differentbusiness models•Crowd-sourced Platforms •Indian managed service providers leverage various workforcemodels with >80% employees from non-metro cities•Model 1: In-house, dedicated workforce, or•Model 2: BPM partnered model •Managed Service Providers•Annotation Platforms as SaaS offerings •India value proposition is driven by decades of service deliveryexperience focused on key pillars - Cost, Infrastructure, Talent,and Innovation •~70% of global players are in the intermediate to advancedphase of labelling maturity Executive Summary (2/2) Indian Players: CurrentMaturity, Challengesand Opportunities Outlook andRecommendations SECTION III SECTION IV •Data annotation industry in India is still gathering pace with ~75%players in the initial to growth phases •2030 Outlook: Data annotation market serviced by India•Can exceed USD 7 Bn. by 2030•Has the potential to engage a workforce of upto 1 Mn. •Challenges restricting the Indian MSPs’ access to markets cover•Data Security & Compliance•Language Barriers across non-English speaking regions•Lack of Cultural Context in labelling for different markets •Roadmap for service providers comprises of strengtheningexisting and developing new capabilities in order to unlockmaximum value •Connectivity and infrastructure related issues are the top COVID-led challenges faced by Indian MSPs •The impact of data annotation on India•Job Growth and Balanced Regional Development•Accelerated AI Readiness •Growth opportunities for MSP players are driven by accessto growing markets,