Product content is one of the most critical factors in the customer buying journey in today's highly competitive digital marketplace. Since...
Why High-Quality Product Content Drives Better Conversions
read more
AI MATURITY JOURNEY
An Efficacy Study in Autonomous Alt Text Generation
By:
Sandeep Dhawan, Senior Executive Vice President, Lumina Datamatics
Sriram Rajagopalan, Director, Presales and New Product Research, Lumina Datamatics
Abstract:
This paper presents an efficacy study of Arty.AI, an AI-powered alt text generation platform developed by Lumina Datamatics. Over three years of operation and more than 2.5 million generated alt texts, the platform has evolved from a standalone application into a high-volume autonomous system with human-in-the-loop (HitL) capabilities for exception handling. Using Levenshtein edit distance as a core performance metric, this study documents how a structured three-step enhancement framework—encompassing an architectural overhaul of AI model, AI-powered quality assurance, and reviewer training—dramatically improved the acceptance rates from 31% to 73%. The findings offer a reproducible blueprint for organizations seeking to implement cost efficient image accessibility in backlist and current publications while maintaining quality and compliance.
Executive Summary
Accessibility is no longer optional— it’s essential for publishers worldwide. Accessible publishing is considered as a strategic priority because of growing regulatory requirements, increasing awareness of inclusive content, and the rising volume of digital assets.
However, one of the most resource-intensive aspects of accessibility compliance is the alt text creation. Traditional manual approaches struggle to scale across large backlists and ongoing content production, resulting in high costs, operational bottlenecks, and inconsistent quality.
Lumina Datamatics’ advanced Arty.AI platform establishes a new approach to accessibility at scale. Through a structured AI maturity journey combining advanced computer vision, large language models, AI-powered quality assurance, and human expertise, Arty.AI improved first-pass alt text acceptance rates from 31% to 73%.
This whitepaper explores the emerging challenges impacting publishers today, introduces a proven framework for AI-powered accessibility, and outlines how organizations can transition from AI-assisted workflows to autonomous accessibility operations.
Authors
Sandeep Dhawan
Senior Executive Vice President, Lumina Datamatics
Sandeep Dhawan heads the Solutions, Transitions, and US, UK Operations for Lumina Datamatics with several decades of experience in driving profitability and operational excellence. He helps the customers by conceptualizing business-impacting solutions which bring value through transformational solutions. His expertise has been around integrated digital-first ecosystems. He has been at the forefront of creating end-to-end solutions to address the customers’ most challenging needs.
Sriram Rajagopalan
Director, Presales and New Product Research, Lumina Datamatics
A skilled solutions architect and strategic thinker, Sriram defines initial product roadmaps, drives R&D of learning management platforms, and delivers thought leadership to represent Lumina Datamatics. He has led presales and new product research, engaging directly with universities, publishers, and scholarly societies to identify needs and design tailored, technology-driven solutions.
