How to Create a Full CRUD Channel using AI

  • Updated on Aug 14, 2025
  • Published on Feb 21, 2025
  • 5 minute(s) read
  • PP
 Prev Next

When ASTRO is activated on the source catalog and the Mapping Method is selected as Full CRUD, users will see an option when creating a new channel:

  • Automatically map nodes and attributes using AI: Switching this field to Yes allows ASTRO (AI) to generate AI recommendations for nodes and attributes.

Launching the (Full CRUD) Channel Using AI

After configuring the channel on the Channel dashboard, you will see that the Launch button will initially be disabled and by hovering over it, you can read the description of the steps:

  1. Syncing the source & target catalogs

  2. Generating Node maps between source & target nodes using AI

  3. Generating Attribute maps between source & target attributes using AI

Once these steps are completed by the AI, the Launch button becomes enabled.

Once the AI is done with all the steps (shown as below), you’ll be able to launch the channel.

Click on the Launch button and you’ll land on the following page with a notification as shown in the screenshot below:

Confidence Score in AI Mapping

The confidence score indicates the AI's level of certainty that a particular mapping is correct. It is represented as a percentage (0% to 100%), with higher scores reflecting greater confidence in accuracy.

For example, when mapping product categories, the AI may assign a 95% confidence score if it strongly believes an item belongs to a specific category, but only 60% if it is less certain.

In PCC, we have set a 90% confidence threshold for automatic mappings. If a mapping falls below this threshold, the AI will provide recommendations instead of applying the mapping directly.

Understanding Node & Attribute Mapping using AI

In this case, similar to any channel, the source catalog is positioned on the left and the target catalog is on the right.

Now that the AI has assigned the generated node and attribute map recommendation, you can see the count at the top of the source catalog along with the AI generated icon.

Note: Default threshold for confidence level score is >=90%.

Let’s look into each highlighted status in detail:

  • Mapped : Displays the total number of nodes auto-mapped by the AI, considering the default threshold set in the backend after channel creation.

  • (Manual): Shows the overall number of nodes that were mapped manually by the user, signifying that the AI did not map any nodes automatically.

  • Provisional : Denotes the total count of nodes that AI recognized as appropriate for the target catalog but were left unmapped due to a higher confidence score threshold for auto-mapping while still providing recommendations.

  • Unmapped : Indicates the number of nodes with AI-generated recommendations that are not mapped, as well as unmapped nodes without recommendations.

Here you can click on the icon or count that functions as a filter for specific node status.

For example, when you click on the iconor count (15), all the 15 nodes that are automatically mapped using AI will be displayed as illustrated below:

From the filtered list, click on any clickable node and you can view icon indicating that the node has been mapped using AI. The connection of the nodes is established via a solid line that links to the corresponding node in target catalog as shown in the image below:

Handling Provisional Mapping

To map nodes with Provisional maps:

  1. Click on the icon or count beside Provisional.

    You’ll see the filtered nodes for provisional map.

  2. Select a node and click on theicon.

    Based on the AI recommendations you can see possible matches with confidence score in dotted lines to the target nodes as illustrated in the image below:

     

  3. Hover over the score to see the reason for the particular mapping suggestion as shown below:

  4. To finalize the mapping; hover over confidence score and click Accept Mapping Recommendation option.

    You’ll see a message as shown below:

  5. Hover the mouse over the gray arrow and it turns blue, revealing a solid line that indicates the node is mapped, as shown in the image below.

    Once the node has been successfully mapped, you will notice that the count in Provisional decreases.

    For instance, if there were 6 provisional maps, after accepting one of the recommended map for a node, it should now show as 5.

    For additional understanding, please refer to the example shown in the images below:

  1. For the node that is mapped, hover on the icon to proceed with attribute mapping following similar process of node mapping.

Attribute Mapping using AI for CRUD Channel

  1. For the node that is mapped, hover on the icon to view the Attribute Mapping Statistics.

    You can view the total attribute count for auto-mapped, unmapped, manual and provisional map as highlighted below.

  1. Click on the Open Attribute Mapping > link on the card.

    You’ll land on the mapping attribute screen.

  2. Similar to node mapping, you can use the count or icon as a filter for the attributes.

  3. To view the mapped attributes, click onor the count besides Mapped.

  4. Select the attribute and hover over the icon to view the attribute mapping done by AI along with the rule that is applied to map source attribute values as is to target attribute.

  5. Clicking on the rule displays its description and allows you to perform actions like editing or deleting it as shown below.

  6. Click Edit, if you would like to edit the existing attribute mapping details.

  7. For Provisional maps, the process is similar to node mapping. To finalize the mapping, hover over the confidence score and click the Accept Mapping Recommendation option.

    Note: Once you confirm the mapping, you cannot reverse it and choose the alternative recommendation

Unmapped Nodes

To map Unmapped Nodes is same as the manual node mapping we used do previously without any AI intervention. For more information refer to Attribute Mapping section.

Refresh AI Mapping Recommendation

If you have new incoming nodes or modified attributes in the catalog, launch the channel and select Refresh AI Mapping Recommendation option under Channel Actions dropdown as shown in the below screenshot:

This option will retrigger the AI for mapping recommendations and you’ll see the following screens:

Now when you land on the Channel Listing page, the Launch button on the channel card will be disabled (as shown in the below image), until the AI has assigned the auto-generated node and attribute map recommendation.

Refresh the channel listing page and you’ll see the Launch button is now enabled, initiate the channel and see the latest AI auto-generated node and attribute mapping recommendations.