In the era of Agentforce and AI, data is the fuel behind intelligent customer experiences. Salesforce Data 360 has transformed how organizations unify and harmonize data using no-code and low-code tools. But as data strategies mature, pro-code developers, data engineers, and data scientists increasingly face complex scenarios that demand the precision and flexibility of custom code. Industry reports show that data professionals spend nearly 40% of their time on data preparation, and 75% say their data pipelines are becoming more complex and brittle.
When handling complex data transformations or unstructured data preparation, standard SQL and drag-and-drop tools aren’t always enough. These scenarios require the power and control of custom code. Today, we’re excited to announce Code Extension, a new capability that lets developers securely execute custom code directly within Data 360.
In our pilot, a customer used Code Extension to read over 139 million rows, apply custom decryption logic, and produce 65 million processed records in just 17 minutes. This is a compute-intensive custom workload that previously required external systems, demonstrating that Data 360 can now handle even the most demanding data engineering workloads.
The Freedom to Build Anything, Right Inside Data 360
Code Extension is designed to make Data 360 the most open and extensible platform for data professionals. Whether you are a Salesforce developer utilizing Agentforce vibes, a data engineer using an IDE like VS Code, or a data scientist working in Jupyter Notebooks, you can now author your Python logic and seamlessly deploy it into Data 360’s secure compute. In the future, we plan to support other languages.
To give you the right tool for the job, Code Extension offers two distinct capabilities:
Code Extension Script: Currently available for Batch Data Transforms, scripts are perfect for complex, multi-step transformations where you need to read data directly from Data 360 Objects and write results back. Some sample use cases:
- Flatten deeply nested XML product catalogs or messy JSON logs into clean, structured tables for analysis.
- Run sophisticated algorithms for prorated billing, risk scoring, or multi-touch attribution models directly in the pipeline.
- Convert raw customer addresses into precise latitude and longitude coordinates using Python libraries to enable logistics and mapping analysis.
- Decode complex binary or hexadecimal telemetry logs from industrial machines into readable formats for equipment monitoring.
Code Extension Function: Starting with chunking in search index, functions allow you to inject specific custom logic into the AI pipeline. Some sample use cases:
- Split documents by logical boundaries like “Speaker” or “Chapter” to prevent AI agents from hallucinating due to fragmented context.
- Split long legal contracts or government policies by specific sections and clauses to ensure compliance agents retrieve exact rulings.
- Apply dynamic chunking rules based on the file type, such as treating spreadsheet rows as individual records while treating presentation slides as holistic chunks.
Built for Developers, Secured by Salesforce
We built Code Extension to fit naturally into the developer workflow while maintaining the strict security standards of Salesforce.
- Developer Experience: Author & Debug code in your local IDE. Test your logic locally before packaging and deploying it to Data 360.
- Safe & Standardized Lifecycle: Adhere to Salesforce DevOps best practices. You can deploy and verify your code in a Data 360 Sandbox first. Once validated, use Data Kits to seamlessly promote your code to Production
- Enterprise-Grade Security: Your code runs in a fully isolated, sandboxed environment to ensure tenant isolation and prevent unauthorized access to your data or infrastructure. We handle the compute provisioning, scaling, and security so you can focus on the logic.
Looking Ahead: Building an Open Platform
We are just getting started. While Code Extension currently empowers you to use Python for Batch Data Transforms and Search Index Chunking, our vision is to open the entire “Ingest to Act” pipeline to your custom logic.
We are actively expanding Code Extension to deliver pro-code extensibility at every stage in Data 360. We also plan to support additional languages, allowing you to leverage your existing expertise and libraries to build in the language you are most comfortable with.
Help us shape the roadmap. Tell us your top use cases by raising a Voice of Customer (VOC) or posting on IdeaExchange so we can prioritize the features where you need the pro code capability.
Get Started with Code Extension
- Learn more: Check out the Code Extension documentation.
- Start Building: Enable Code Extension in feature manager, navigate to the Code Extension tab in your Data 360 org to deploy your Code Extension today.
