Resize image to 1920x680px before upload

The illicit massage industry (IMI) is arguably one of the largest and most networked forms of human trafficking in the United States. An illicit massage business (IMB) is an establishment that puts on the façade of a legitimate massage business to facilitate commercial sex services. As of 2023, our client, a major US-based non-profit organization, had assessed that there were over 13,000 IMBs operating in all 50 states – and growing. This national industry generates over $5 billion a year in illicit revenue. Traffickers are organized, coordinated, and effective. Whether in big cities or small towns, they are enabling and profiting off the exploitation and trafficking of women across the country.  

To combat human trafficking, our client uses data scraping and analysis techniques to identify illicit massage businesses (“IMB”) based on web-based listings and publicly available data. The ability to scale their data intake and automate identification of unique targets presents a substantial challenge to the business. To solve this challenge, they partnered with Slalom to complete an assessment of current state technologies and identify opportunities to address key intelligence gaps and develop a minimum viable product (“MVP”).  

The MVP, built on AWS-native services including EC2, S3, CloudWatch and Lambda, provided an ingestion pipeline based on current data, to identify duplicate IMBs using unique business attributes and classify business records. This also provided an important framework for further integration with additional data sources, allowing the organization to both ingest more sources for known IMBs while also maintaining a scalable process for evaluating and targeting these businesses. As part of the roadmap, Slalom also developed a process to deduplicate records of illicit businesses, allowing improve efficiency and reduction of redundant and manual efforts. 

Together we developed a unique ingestion processing pipeline, using attributes of a business record to identify if a business has previously been identified. This process uses phonetic matching on business names, phone number and address values, and classification scoring algorithms to develop an 80-85%+ identification accuracy. Using the existing 40k records in the organization’s database, this process was able to identify and remove ~9k records (25%) and highlight overlap between disparate listing sites. With this solution now part of their production workload process, our client can better focus on IMB targets while also providing the foundation to now support additional data intake.  

This solution helps establish when a given business was first discovered, how many times it has been identified and from which data sources, and how active the business is. Combining this list of unique businesses and activity, the organization can drastically reduce the number of targets and better focus their team. This allows them to partner more strategically with law enforcement agencies to shut down these illicit businesses faster and more efficiently.  

Interested in learning how your organization solve business challenges with Amazon Web Services? Looking for a trusted partner to help you along the way? Reach out to Slalom to learn more today.