The Network Partners with Slalom and Uses AWS Machine Learning and AI to Help Combat Trafficking in the Illicit Massage Industry

The Network is a data-driven counter-trafficking organization, focused on eradicating the illicit massage industry. To automate and scale its data scraping and analysis capabilities, the organization worked with AWS Partner Slalom to build a data ingestion pipeline using Amazon Web Services (AWS) technologies. Automating this process allows The Network to strategically focus people resources and avoid duplicative efforts. As a result, The Network identified and removed 9,000 duplicate records, achieved 85 percent model accuracy, and automatically classified and analyzed half a million records to better support law enforcement agencies in shutting down illicit massage businesses.

One of the largest forms of human trafficking in the United States is the illicit massage industry. There are more than 10,000 illicit massage businesses (IMBs) that pose as legitimate operations to facilitate illegal sex services. Dedicated to eradicating this growing problem, The Network fights human trafficking by scraping and analyzing web-based listings, ads, customer reviews, and other publicly available data to enable state and local law enforcement to dismantle IMBs.

To identify more IMBs, The Network wanted to expand and automate its data collection, deduplication, and classification processes, which were time-consuming and often manual. “There are a lot of nuanced steps associated with data scraping and analysis, and we really wanted to automate the entire process instead of relying on our own data pipelines, which weren’t as efficient,” says Louis Wilbrink, chief technology officer at The Network. The organization also sought to expand the number of online data sources it could collect. “We wanted to be able to scale more easily to capture additional data, from more sources, so our conclusions were more informed,” says Wilbrink. “We need to be as confident as we can when we speak with our law enforcement partners about our findings.”

In 2023, The Network engaged Slalom, a business consulting technology company and AWS Partner, to help address its automation and scalability requirements. Slalom and Amazon Web Services (AWS) led a full-day workshop with The Network, identifying challenges and strategizing new technology solutions. “The Network wanted to deduplicate data that they see across multiple sources to ensure they’re not targeting the same business because an address is slightly different, for example,” says James Sturges, senior director of software engineering at Slalom. “We also wanted to see if we could integrate things like sentiment analysis as a way to diversify data sources. We came out of the workshop with a tangible roadmap for building a new solution.”

Following the workshop, Slalom worked closely with The Network to design and develop an AWS-based data ingestion pipeline that identifies duplicate IMBs using unique business attributes and business records. The pipeline’s process uses a combination of phonetic matching on business names, phone numbers and address values, and classification scoring algorithms. “Many services and businesses are listed in different ways, and that’s a unique challenge because there's no identifier like a tax ID for illicit businesses,” says Sturges.

The solution uses Amazon Bedrock large language models (LLMs) and Amazon SageMaker models to vectorize content and compare it to known illicit activity keywords and phrases. The content is stored in Amazon Simple Storage Service (Amazon S3) buckets. Additionally, The Network uses Amazon Elastic Compute Cloud (Amazon EC2) to host data scraping and processing tools, and AWS Lambda and AWS Step Functions to orchestrate the data deduplication process and a multi-step workflow for cleaning, validating, and processing business records.

With the ability to use unique business record attributes to determine whether an IMB has previously been identified, The Network found and removed 9,000 duplicate records—25 percent of its database. The identification model had an 85 percent accuracy rate. “We are seeing remarkable results through the combination of AWS's advanced machine learning capabilities and Slalom's strategic consulting approach,” says Wilbrink. “The ability to analyze and process massive amounts of data while maintaining accuracy demonstrates how technology can be used to make a meaningful impact in critical initiatives. Together with AWS, we're not only building solutions—we’re enabling organizations to transform their operations and achieve meaningful outcomes.”

The Network is using its data processing pipeline to establish when an IMB was first discovered, how many times it has been identified and from what data sources, and how active the business is. The organization can also scale its pipeline to process more IMB records. “We have a foundation now to support additional data intake, and the machine learning algorithm automatically classifies the data,” says Wilbrink. “As a result, we can quickly scale the number of classified reviews in our database. We have half a million reviews, and we couldn’t analyze those manually. We’d have to hire a hundred people working 24x7 for months to achieve that, and this solution allows us to do it in a fraction of the time and with better consistency.”

The Network now has a framework for further integration with other data sources, allowing the organization to ingest more sources for known IMBs while also maintaining a scalable process for evaluating and targeting businesses. With more accurate data, The Network can improve access to data for its partners and better support them in shutting down IMBs. Wilbrink says, “With Slalom and AWS, we have much more confidence in the data, which improves our partnerships with local law enforcement agencies.”

The Network, based in Arlington, Virginia, is a data-driven counter-trafficking organization that enables law enforcement partners to measurably reduce exploitation at scale. Its diverse team uses skills developed in the intelligence community, data analytics, financial services industry, and victim service provision.