Deploy Externally Hosted AI Models in Magic ETL

We are excited to announce our latest Domo.AI innovation that makes third-party AI models more accessible to everyone—transforming data prep and analysis. Our new AI Inference tile allows you to deploy any of your eligible externally hosted models, including those from Databricks and Amazon Bedrock, within Magic ETL using Domo’s ML connectors. Please note that currently, models must have CSV input and output formats, but this will expand to include JSON-based models in the near future.

By leveraging AI with their own business context, low-code users can now utilize the brilliant models developed by your high-code talent to identify new trends, accelerate analysis, and quickly share insights—no advanced coding skills required! Sounds game-changing, right? Now, let’s bring it all to life. Imagine you have valuable sales data, and you’d like to use it to identify new market opportunities, optimize resource allocation, and accurately predict future revenue. Your data team developed a custom model for forecasting based on your company’s needs, but it lives in Databricks—a platform you’re unfamiliar with. With this new feature, you can seamlessly integrate that model into Domo and run your data against it to create the forecast you need to optimize business strategies and fuel more confident decision-making—with precision.

Simply provide the data you wish to analyze and specify where the results of the model's insights will be stored or utilized. Once everything is configured, save and execute the dataflow, allowing the model to analyze your data and deliver valuable insights based on its predictions.

With this new feature, you can:

• Streamline the data preparation and analysis process, making it even more efficient and accessible to everyone.
• Quickly identify trends and patterns, accurately predict future outcomes, and surface intelligent insights, faster.
• Enable business users to seamlessly disseminate data, visualizations, and insights across your organization.