Roboflow + Salesforce
Turn a Workflow detection into a Salesforce Work Order or Case with the evidence image attached, using one custom block you configure in your browser.
How to connect Salesforce to Roboflow
5 stepsConfigure the block
Open the configurator and enter your Salesforce instance URL, site, crew and SLA hours for each severity. It writes the block code for that setup.
Add a Custom Python Block
In the Workflow builder, add a Custom Python Block after your model and visualization blocks, choose Edit Code and paste the generated code.
Declare and wire the inputs
Declare the inputs and outputs with names that match the code. Connect image to the visualization output, predictions to the model, and access_token to a workflow input named salesforce_token.
Test with a dry run
Set destination to dry_run, run the Workflow on a test image and check the payload output before anything reaches Salesforce.
Go live
Switch destination to salesforce_work_order or salesforce_case, run the Workflow with your token and open the record link in the result.
What teams build with Salesforce
Open a work order when a defect is found
File a maintenance Work Order with the frame attached when a model spots a fault on a production line.
Log a safety incident as a Case
Create a Case with the evidence frame when a Workflow sees a safety problem, so it is tracked like any other incident.
Send the alert to another system
Use the webhook destination to post the same payload to any system that accepts webhooks.
Works with
| Product | How Roboflow connects |
|---|---|
| Salesforce Field Service | Work Order destination |
| Salesforce Service Cloud | Case destination |
Questions
No. Work Orders come from Field Service. Without it, choose the Case destination and the rest of the setup stays the same.
In a workflow input you pass at run time, never in the Workflow definition. For production, use a Connected App with the client credentials flow and a user that can create Work Orders or Cases and files.
Object detection, instance segmentation, keypoint and classification models, from any model block in the Workflow.
The Workflow keeps running. The block reports the failure in its error_status and message outputs instead of stopping.
Yes. Custom Python Blocks run on the Serverless Hosted API, Dedicated Deployments and self-hosted Inference.
Why are companies choosing Roboflow?
The end-to-end platform for vision AI: create datasets, train models and deploy to production.
Flexible deployment modes
Move from a scalable cloud API to edge hardware like NVIDIA Jetson without managing complex dependencies.
Integrated workflow builder
Combine multiple AI models, custom business logic and third-party integrations into one production-ready pipeline in a low-code editor.
Model monitoring and alerts
Dashboards track confidence, inference time and request volume for your deployed models, and email alerts tell your team when confidence suddenly drops.
