The challenge
The compromised PII was saved in numerous formats: Word documents, PDFs, emails, ZIP files, JPEGs, PNGs, and PowerPoint presentations. It spanned various sensitivity levels, from high-risk data such as payroll information to lower-sensitivity data such as names and countries of residence. The client’s priority was to notify those individuals facing a higher risk. This required understanding the nature of each person’s exposure.
Our solution: A custom-built AI platform for extraction and analysis
Reed Smith Legal Solutions developed a custom AI solution engineered to read and interpret an extraordinary range of file formats securely within a controlled environment. The team first singled out a subset of 10,000 sensitive documents (such as passports and ID photos) and tagged them as potentially containing PII.
It then deployed an AI solution to scan and extract data from the subset – validating names, passport numbers, and expiration dates – to get the full list of individuals. Traditional review methods for this dataset were estimated to require 600 to 800 hours, but the team completed its AI-enhanced review in just two days with fewer errors. The AI could even filter out visual obstructions, such as the holograms that overlay passport pages, enabling it to read the underlying text just as a human reviewer would, but at vastly greater speed.
All results were organized into a searchable database, categorized by urgency and sensitivity level: high, medium, and low. The legal team could then query the data instantly, generate statistics, and obtain direct answers for regulatory discussions.
The results: Six months faster resolution
With the data set in hand, the legal team provided regulators with qualitative analysis and direct answers about the scope and nature of the breach. The client completed the impact analysis of the breach as much as six months earlier than would have been possible with a traditional approach – and was able to notify affected individuals within regulatory deadlines.
