Visual AI Workflows

Workflows & Tutorials
October 1, 2025
|
Aria Bennett

In this guide, we’ll walk you through how to build, manage, and monitor Visual AI Workflows with SunSummit. Using our Workflow Builder, you can easily design and automate end-to-end workflows without needing to write complex scripts.

Visual AI workflows let you connect datasets, models, and processes in a simple, intuitive interface. Whether you're cleaning data, training models, or deploying predictions, SunSummit helps you automate the entire lifecycle.

1. What is a Visual AI Workflow?

A Visual AI Workflow allows you to design and automate machine learning processes through a drag-and-drop interface. It gives you the ability to create data pipelines without writing code, while still maintaining flexibility and control over each step in your workflow.

With SunSummit's Workflow Builder, you can:

  • Create pipelines visually — no need to code.
  • Trigger actions automatically — set rules for when specific tasks should be executed.
  • Track data lineage — see how data moves across various steps.
  • Monitor pipeline health — stay informed about performance, latency, and potential issues.

2. Getting Started with Workflow Builder

To create your first visual AI workflow, follow these simple steps:

  1. Open your SunSummit dashboard.
  2. Navigate to "Workflow Builder".
  3. Click "Create New Workflow".
  4. Drag and drop elements to design your workflow.

You can add steps like:

  • Data Ingest (e.g., from S3 or databases)
  • Data Processing (e.g., cleaning, normalization)
  • Model Training (e.g., start training using a pre-built model or custom model)
  • Prediction (e.g., batch predictions or real-time scoring)

3. Example Visual AI Workflow

Here’s an example of a Visual AI Workflow that cleans data, trains a model, and deploys predictions:

1. Data Ingest:
First, ingest raw data from your data lake or cloud storage.‍

pipeline:
  name: data-preprocessing
  steps:
    - input: s3://raw-data

‍‍2. Data Transformation:
Clean and normalize the data.

pipeline:
  steps:
    - transform: normalize
    - transform: dedupe

‍3. Model Training:
After preprocessing, use a pre-trained model or create a custom model.

pipeline:
  steps:
    - model: train
    - output: model-v1.0

‍4. Deployment:
Deploy the model to production for real-time or batch processing.‍

pipeline:
  steps:
    - deploy: api
    - output: s3://model-outputs

‍4. Advanced Workflow Features

Once you're comfortable with basic workflows, you can dive into advanced features to take your workflows to the next level.

  • Conditional Logic:
    Use conditions to run specific steps based on certain criteria. For example, run step 2 only if step 1 completes successfully.
pipeline:
  steps:
    - input: s3://raw-data
    - condition: if "data_quality_check == pass"
    - transform: clean
  • ‍‍Event-Driven Pipelines:
    Trigger workflows based on events such as file uploads, scheduled times, or system changes.
sunsummit workflow trigger --event new-data-upload
  • ‍Monitoring & Alerts:
    Set up real-time monitoring and alerts for workflow issues such as failures, latency, or drift.
sunsummit monitor workflow --alerts "failures, latency, drift"

5. Monitoring Your Workflow

Once your workflow is up and running, SunSummit allows you to monitor its health and performance in real-time.

You can view:

  • Pipeline status: Check if each step is successful or failed.
  • Data quality metrics: Track the quality of incoming data.
  • Model performance: See how well your models are performing.

Example: Check workflow status:

sunsummit workflow status my-workflow

Example: Check data quality metrics:

sunsummit metrics data-quality --workflow my-workflow

6. Next Steps

Now that you’ve created and monitored your first visual AI workflow, here are some next steps:

  • Explore more advanced workflows: Create more complex pipelines and automations.
  • Integrate external systems: Connect SunSummit to cloud storage, databases, and other tools.
  • Share your workflows: Collaborate with teammates or stakeholders on shared projects.

Continue to explore the documentation for additional guides, best practices, and advanced configurations.

Conclusion

With SunSummit’s Visual AI Workflow builder, you can streamline your entire ML pipeline, from data collection and cleaning to model deployment and monitoring. This guide covered the basics of getting started with visual workflows, but there’s much more to explore. Keep experimenting and building efficient workflows with SunSummit.