PUBLISHED DATE: 2026-04-29 09:46:36
VIDEO TRANSCRIPT:
SPEAKER A:
Hi, I am Tanu, VP of Engineering at Publicis Sapient. Migrating legacy systems can take a very, very long time. Today, I am going to show you how we use Sapien Slingshot to convert code from old COBOL applications and platforms into new code, shaving years off the migration process. Here is how we use Slingshot to convert an outdated COBOL-based application into a cloud-native, modernized Java microservices-based application. Here is the Sapient Slingshot Agent Portal. We'll start by selecting what code we are converting and the outputs we want. And now we'll upload our legacy COBOL code. First, Slingshot will analyze the code. Real-world COBOL codebases can be too large to fit effectively in LLM context windows. For Slingshot intelligently breaks this into chunks using logical, abstract syntax tree parsing and code adjacency methods. COBOL files can also have multi-level dependencies. So Slingshot uses a hybrid rack with hierarchical vector and graph stores and journey mapping to discover dependency trees. Slingshot then traces the data entities and logical flows to create dependency diagrams. Here our extendable code to speak agents are scanning the COBOL code using various MCB tools to convert it into comprehensive business and functional specifications, outlining the inputs and outputs, process flow and relevant data structures. This step gives us a detailed understanding of what needs to be updated with up to 99% accuracy. Now that we have our comprehensive specifications, including acceptance criteria, validation rules, and detailed journey context, we move on to the modernization process. Here, I can use Sapien Slingshot to convert these business requirements and specifications into Java code using our Spectre code agent that brings in a large set of prompt libraries and guidelines for cloud-native architectures, curated by public Sapien experts. Enterprises can also inject additional business and architecture guidelines that are relevant to current and future needs. This will not only improve performance, but also ensures consistency and compatibility with cloud native architectures. Security is critical in modernization. The security process ensures our updated Java code follows best practices for authentication, data protection, and compliance. It automatically applies security patches and flags any potential risks. It also identifies unpatched dependencies, redundancies and eliminate stale unused code. With the code modernized and secured, we can move on to testing. Sapient Slingshot includes agent-based testing tools that validate the functionality, performance and reliability of the new Java application. Now we are ready to deploy. With just one click we can push the modernized Java application to the cloud using our deployment agents. And that's it. We have taken legacy COBOL code and transformed it into a modern, secure and scalable Java solution that works on the cloud. Now you can more easily do things like launch new product features without old code holding you back. Thanks for watching. It's unlicensed prompting limited, sheesh it says. Okay.