Branching out into any new vertical or service line can be a daunting undertaking for any business. Success in potentially attractive and high-growth markets may seem unattainable, or too risky to warrant an attempt, particularly if there is no obvious path in.
One such market is data-for-AI, a new and highly specialized sub-industry that has emerged within the high-growth artificial intelligence (AI) industry, with businesses competing to provide large volumes of quality, labeled data to companies that build AI solutions.
Market leader Appen has found long-lasting favor with the stock markets since its debut more than five years ago. The company’s performance throughout 2020 offers potential market entrants the prospect of a (Covid-19) resilient offering.
Appen aside, the data-for-AI market has gained increasing visibility among the “core” landscape of language service providers (LSPs) in recent years. Interest has only intensified with the sale of Lionbridge’s sizable data division, Lionbridge AI, for just shy of USD 1bn in late 2020.
What’s more, an increasing number of LSPs from across the revenue spectrum are building or acquiring capabilities in data-for-AI services. Although it will never be a lift-and-shift endeavor, the data-for-AI market represents a significant opportunity, and one that LSPs are uniquely positioned to pursue.
As multi-vertical experts, many LSPs already have touch points with end-customer sectors that are beginning to apply AI in their own specific ways. Here, we lift the curtain on some of these key use cases for AI, and outline how an LSP’s customers across a range of sectors leverage text, image, and speech data for use in training AI applications.
6 Sector-Specific Tips for LSPs
- IT / Big Tech
Use cases in AI in big tech and AI are vast and include: computer vision capabilities, machine translation (MT), semantic search, text-to-speech, virtual assistants and chatbots, and speech recognition.
LSP insight: As big tech companies are on the cutting edge of new AI capabilities, clients may have more specific (or unique) data needs compared to typical business cases. Consultation services to establish use case and data strategy may be useful.
In the highly-regulated sectors of finance and banking, AI applications are used for fraud detection and building risk-management models, as well as in text-to-speech systems and equipping chatbots for customer interactions. In investment banking, AI and machine learning are used in algorithms to assist trading and for stock market predictions.
LSP insight: Data security is a key consideration for financial clients. Companies may require onsite, security-cleared data annotation and transcription. The finance industry is a rich source of data for machine learning, having maintained financial quantitative records for decades.
The primary use cases for AI in the automotive sector are speech recognition (for driver assistance) and autonomous vehicle capabilities, including object and hazard recognition.
LSP insight: Autonomous vehicle capabilities are built on algorithms that are fed labeled image, video, and sensor fusion data. Synthetic and modified image data now augment manually-labeled data.
AI assists in a broad range of healthcare tasks. In medical imaging and diagnostic healthcare, AI is used in diagnostic error prevention and providing medical imaging insights. In the area of patient care, AI helps in the analysis of patient data. Additional patient-care use cases include pregnancy management (e.g., AI-backed ultrasound simulation platforms), real-time prioritization of patients or cases, and automated prescriptions.
Within healthcare R&D, AI can be applied in the area of drug discovery, where it helps predict the properties of a potential compound, alleviating repetitive manual lab tasks, and in genetics (for analytics and editing), as well as in comparing the effectiveness of devices and drugs.
LSP insight: Healthcare data labeling is at the high end of the subject-matter expertise spectrum; for example, identifying tumours in MRI scans requires special training. While healthcare is rich with decades of patient data and diagnostic images, repurposing this data for machine learning requires extensive data curation, cleaning, and data management, in addition to labeling.
- Retail & E-commerce
The primary use for AI in retail and e-commerce relates to behavior analysis and customer experience. For example, AI can help in search relevance, chatbot training, sentiment analysis, and user recommendations, as well as in supply chain planning, customer engagement and demand forecasting. In addition, AI assists in categorizing products, onsite searches, social media analytics, and smart technology (i.e., the Internet of Things).
LSP Insight: Search relevance and sentiment analysis labeling are increasingly in demand.
Governments leverage the use of AI in a variety of ways, many of them dealing with sensitive data. Use cases include social media monitoring, computer vision, extracting information to aid surveillance and monitoring of national security threats, voice recognition, national emergency response, chatbots and virtual assistants for public interactions with government, contributing to public policy objectives, and research and development.
LSP Insight: Data security is a key consideration and onsite security-cleared data annotation and transcription may be required. Government adoption of AI is uneven and is generally lagging behind the private sector.
For a deep-dive into this highly attractive niche, download Slator’s full, 44-page report on the data-for-AI market. The report guides LSPs on how to enter and scale in the fast-growing market of creating, collecting, and annotating data for artificial intelligence applications, featuring five case studies of LSPs that serve data-for-AI customers, one case of a publisher-turned-data-for-AI provider, and much more.