Skip to main content

DataOps.live Assist — your AI-Powered Copilot

Feature release status badge: PriPrev
PriPrev
DataOps copilot for the data product platform

DataOps.live Assist is an AI-powered copilot that improves your user experience on the data product platform. It enhances the workflow for every team member, regardless of their role. Use it to massively increase each team member's productivity and do more with the same.

  • As a data engineer, use it during your development cycle.
  • As a data product owner, use it during review and release management sessions.
  • As a support engineer, use it during incident resolution as guided action.

Assist for data engineers

Assist supports data engineers during the development of MATE dbt models. The main use cases are:

  • Suggest models from natural language inputs
  • Describe SQL or Python that is missing documentation
  • Explain 1000+ lines of complex SQL

Overall, you can enjoy three times better productivity.

Assist for data product owners

Assist automatically summarizes and describes complex change requests, helping data product owners quickly understand and approve merge requests. This speeds up the process, meeting business needs faster and reducing time to production by 50%.

Assist for support engineers

Assist analyzes data pipeline failures, identifies the root cause, and provides natural language steps to resolve the incident.

With less back-and-forth communication in your team and with Dataops.live Support, experience a 10X faster resolution time for incidents.

Developing your SQL models

Dataops.live Assist provides data engineers with integrated tools to aid in the creation of MATE models. You can access these tools through the DevReady command palette within the DataOps development environment.

Assist Command Palette !!shadow!!

Opting into DataOps.live Assist on DevReady

To access Assist on DevReady, do the following:

  1. Set DATAOPS_FEATURE_ASSIST = true as an environment variable in GitPod:

    1. In DevReady, go to Gitpod → User Settings → Variables (https://develop.dataops.live/user/variables).
    2. Add DATAOPS_FEATURE_ASSIST = true for your development workspace.

    Define environment variable to access Assist !!shadow!!

    Using scope */** makes the environment variable available to all user workspaces.

    Alternatively, run gp env DATAOPS_FEATURE_ASSIST=true in the DevReady terminal. Note this will make the variable available only to the project you are currently working on. For example, the scope will be automatically set to my-group/my-sub-group/my-project.

  2. Restart your workspace and you will have access to DataOps.live Assist.

Creating new models

Create a new Model suggests MATE dbt models from your natural language inputs.

Based on your prompt, you will have an initial <model>.sql created in seconds. In addition, the corresponding <model>.yml containing the model documentation will be generated for you.

  1. Run Create a new Model from the command palette in DevReady.
  2. Enter your query description and set a path for the model you want to generate.
  3. Select the sources corresponding to your query and click Create to generate the SQL model and its YAML automatically.
  4. Run your query and check the results in the console.

Documenting models

Document this model describes your SQL or YAML model that is missing documentation.

Given a SQL model, the tool generates a <model>.yml which is fully documented with model and column descriptions.

Given a YAML file that is missing descriptions or column information, the tool uses the matching <model>.sql file to document the descriptions in the <model>.yml file.

You can review and choose which changes to commit to Git.

  1. Open the SQL model or YAML file without a description.
  2. Run Document this model from the command palette in DevReady or by clicking the pencil icon in the top right of the editor pane.

Explaining models

Assist explains your SQL model and generates a summary to help you start your development work.

This is helpful for long and complicated SQL queries, as it can be difficult to understand 1,000+ lines of code.

  1. Open the SQL model you need an explanation for.
  2. Run Explain this SQL from the command palette in DevReady or by clicking the icon in the top right of the editor pane.

A summary and explanation are generated for the open model.

Giving feedback

You can provide feedback on Dataops.live Assist either through the pop-up that appears while using any of its tools or at any time from the left sidebar.

  1. Click the Dataops icon and go to Assist Feedback.
  2. Select the tool you wish to provide feedback for and share your insights.
  3. Click Submit.

Speeding up merge request review and approval

Once a merge request is created, Dataops.live Assist automatically generates an insightful MR summary. It makes reviewing easier by giving a short and informative summary of the changes in an MR.

All private preview users have assist enabled by default, enhancing their experience with seamless MR summaries. When you create an MR, you'll see a new activity in the MR overview with this message:

Initial DataOps AI assistant message for merge requests !!shadow!!

Assist thoroughly analyzes the MR changes. After reviewing, It records a message under Activity. The message includes a summary and detailed explanation of the changes.

Subsequent DataOps AI assistant message for merge requests !!shadow!!

Opting out from DataOps.live Assist

While Dataops.live Assist is beneficial for most users, your preference may still vary. If you want to opt out of Assist, email us at privatepreview@dataops.live for assistance.