Case study

Genpact's AI-powered leap into the future: Transforming background check validation

How Genpact used Amazon Bedrock to accelerate and improve employee screening audits

Who we worked with

Internal stakeholders at Genpact, a leading professional services and solutions firm.

What the company needed

To automate the process of validating employee screening reports to support rapid assignment to any of our more than 600 clients worldwide
To improve compliance with clients' employee background check requirements

How we helped

Developed an end-to-end gen AI solution, powered by Amazon Bedrock, to automate validation of background screening reports for any of our more than 125,000 employees and be compliant with varied client requirements

What the company got

A scalable gen AI solution for background check validations using Amazon Bedrock
Up to 90% reduction in manual verification, enabling associates to focus on other key areas like record to report
Up to 80% machine accuracy achieved for employee background check verification against client requirements
Increased contractual compliance with 100% of associates covered, helping Genpact meet client requirements and improve risk mitigation

THE CHALLENGE

Automating the employee screening verification process

At Genpact, we take client requirements for employee background checks very seriously. Not only do we conduct thorough background checks – internally or through partners – for new employees or employees who switch roles internally, but we also have a robust process in place to validate background check reports.

However, validating these employee background check reports against client requirements used to be a manual process. The information for these audits would often lie across several individual files, necessitating manual searches through structured and unstructured data in multiple files. This meant that Genpact's audit staff was burdened with cumbersome routine tasks that kept them away from higher-value work.

Our IT team took up the challenge of fast-tracking the development of an automated, AI-led interface to replace the time-consuming legacy process.

Key considerations included:

Automating the validation processes
Increasing the accuracy of the results
Verifying all background reports rather than just performing sample checks

THE SOLUTION

Harnessing the power of AI to accelerate background check validation

As a first step, Genpact's developers needed to create a roadmap for consolidating employee files and client compliance requirements into a central data repository. Around the same time, Amazon Web Services (AWS) introduced Amazon Bedrock, a fully managed service offering a broad set of capabilities for building generative AI applications. Genpact, which was already running multiple workloads on AWS and using Amazon SageMaker for machine learning modeling, immediately began looking into the new service. With assistance from AWS experts, the IT team zeroed in on Amazon Bedrock and Anthropic's gen AI assistant, Claude, to develop a generative AI application to automate validation of employee background check reports.

We created a bot to consolidate data from multiple background check reports into a single format and implemented an AI-powered solution to read and interpret compliance requirements. This enabled us to replace our manual system for determining compliance with an automated, AI-enabled process.

Our team created a new Amazon Bedrock model that enables Claude to check employee background documents against client vetting requirements. Encrypted data is stored in the cloud and processed using Amazon Glue, a serverless data integration service that uses parallel processing to handle large workloads.

Background check audits are initiated by our internal audit teams using a customized Microsoft Azure interface. Plus, we built an open API gateway to process verification requests. In addition to screening for clients, the system automates background check compliance audits for our own new hires.

THE IMPACT

More accurate background check validation in half the time

The solution, powered by generative AI and built on Amazon Bedrock, expedites the team's ability to sort through structured and unstructured data to verify employee compliance. The new interface can accurately filter data on more than 3,000 employees a month to ensure they meet the unique contract requirements of each client.

Our standardized process saves time and reduces the likelihood of human bias or manual processing errors. Offloading time-consuming manual processing also increases efficiency. At the same time, keeping humans in the loop for final verification helps boost compliance and mitigate risk.

Here's a snapshot of the concrete outcomes the transformation delivered:

Processing time was cut in half, from 5 days to 2.5 days
Manual effort was reduced by approximately 75%
Up to 80% accuracy was achieved using AI to verify against client requirements
The need for manual verification was restricted to just 20% of screening requests
We are now able to validate 100% of our background checks, improving contractual compliance

Embracing the power of generative AI for background check verification has propelled our internal audit team into a new era of efficiency and excellence. They are now able to focus on more strategic initiatives, like record to report, that drive our company forward.

By enabling comprehensive compliance checks for every employee assigned to client accounts, we've elevated our competitive edge. This meticulous attention to detail safeguards our clients and reinforces our position as a trusted partner. Most importantly, by harnessing the potential of AI to solve real business problems, we move closer to our vision of constantly innovating to create a more efficient, trustworthy, and client-focused organization.

Insights section: