Building a Sub-Metric Generator using AI

DataBee uses AI to build a sub-metric generator, a configuration object for our data streaming infrastructure, that gathers metrics from third-party systems. Every time a metric must be updated the associated sub-metric generators are reused.  

Limitations

DataBee requires network connectivity to third-party data sources in order to build a sub-metric generator. The third party data source must also have publicly accessible API documentation. If API documentation exists but is not publicly available contact DataBee Support for assistance.

Step 1: Connect data source to the sub-metric

Connect a data source to the sub-metric using these steps:

• Locate: Provide an endpoint, documentation link, or OpenAPI specification.

• Authenticate: Enter the required credentials. If you do not have them, select Draft an email for me to request access.

• Confirm Mapping: Review and confirm that the retrieved data is mapped correctly to the submetric.

• Connect: Complete the connection.

Use the following options to manage the current conversation:

• Minimize: Hide the panel while the search continues.

• Park: Save the current conversation so you can return to it later.

• Start Over: End the current conversation and begin a new one.

Select from pre-existing Connectors or Research a new connector. DataBee will use online documentation and its knowledge base to identify supported authentication protocols and the require parameters for that product’s implementation of the protocol. The form below is an example and available required fields will vary by product.

DataBee can also write an email for you to send to the product owner requesting a service account with the required details.

Step 2: Build the Sub-Metric Generator

After selecting or creating a connector DataBee uses the connector to access the third-party data source. It uses its knowledge base and online research skills to find appropriate API endpoints to use based on the description of the sub-metric.

Once the sub-metric generator is built the sub-metric value is calculated and data samples are retrieved for review.

Step 3: Review the sample Preview records and Refine

Select Preview Records to see a small number of records included in the sub-metric value. Investigate available fields and if the samples include unintended items click Refine and provide updated instructions.

Step 4: Refine the Sub-Metric

At any point in the process you can Click Minimize to reduce the chat window to a small summary bar in the lower right-hand side of the screen.