Getting Started With Assistants

Updated: August 28, 2026
Contents

    The Onboarding Assistant guides you through turning a catalog spreadsheet into live Products and Items in CloudBlue Connect®. This walkthrough follows the five stages of the wizard: Source → Mapping → Validation → Create → Results — so you know exactly what to expect and what to check at each step.

    Before you begin: Have your catalog file ready as a single Excel or CSV file. See Preparing The Catalog File for formatting guidance that will make the AI’s mapping more accurate on the first pass.

    Step 1: Source — Upload your catalog file

    1. From the CloudBlue Connect® navigation menu, open Assistants > Onboarding Assistant.
    2. Select New Run.
    3. Drag your file into the upload area, or click Browse to select it from your computer. Supported formats are .xlsx, .xls, and .csv.
    4. The Assistant runs a quick check on file type and size before accepting it. If the file is rejected, you’ll see the specific reason (for example, an unsupported format or a file that’s too large) rather than a generic error.
    5. Click Start Run.

    At this point the Assistant reads your file and begins generating a mapping plan. This typically takes a few moments — you don’t need to keep the tab open; the run continues in the background and you can return to it from Run History.

    Note: Starting a run does not create anything in your catalog yet. Nothing is written to CloudBlue Connect® until you explicitly confirm in the Create step.

    Step 2: Mapping — Review how your data was interpreted

    Once the plan is ready, you’ll land on the Mapping screen. This is the most important step to review carefully, since it determines how your spreadsheet rows become Products and Items.

    What you’ll see:

    • A tabular view of your source columns alongside the Connect fields the Assistant thinks they correspond to (for example, a column named SkuTitle mapped to the Item’s name field).
    • A confidence indicator next to each mapping. High-confidence mappings are inferred with strong signals from your column names and data; lower-confidence mappings are flagged for your attention.
    • Grouping logic, showing which column the Assistant used to tell Products apart from one another, and which column it used to identify individual Items within a Product.
    • Inline definitions for Connect concepts like Products vs. Items, shown in context next to the fields they apply to, so you don’t need to leave the page to look them up.

    What to do:

    1. Review each row of the mapping table. Accept mappings that look correct.
    2. For any mapping marked Needs your review, either approve the Assistant’s suggestion, choose a different source column from the dropdown, or enter a fixed value if the field should be the same for every record.
    3. Pay particular attention to fields marked Required — these must be mapped (from a column or a constant value) before you can continue.
    4. When you’re satisfied, click Continue to Validation.

    Tip: You can go back and adjust mappings later if validation surfaces an issue that traces back to a mapping choice.

    Step 3: Validation — Resolve data issues before anything is created

    The Assistant checks your mapped data against Connect’s requirements and flags anything that would prevent a successful Product or Item creation.

    Common validation categories:

    CategoryWhat it meansWhat to do
    Missing required dataA required field has no value for one or more rowsFill in the missing values in your source file and re-upload, or map a different column
    Type mismatchA field expects a specific format (for example, a number or a fixed set of values) and the source data doesn’t matchCorrect the data type in your source file, or adjust the mapping
    Possible duplicatesTwo or more rows appear to describe the same Product or ItemConfirm whether this is intentional, or remove/merge the duplicate rows

    Where possible, you can fix simple issues directly on the Validation screen without leaving the run. For larger data-quality problems, it’s usually faster to correct the source file and start a new run.

    Click Continue to Create once all required issues are resolved. (Warnings that don’t block creation are shown but won’t stop you from proceeding.)

    Step 4: Create — Confirm what will be created

    Before anything is written to your catalog, the Assistant shows a confirmation screen summarizing:

    • The total number of Products and Items that will be created.
    • Any constants or default values that will be applied (for example, a default subscription period).
    • Any assumptions the Assistant made that you should be aware of.

    Review this summary carefully — it’s your last chance to catch something before records are created. When you’re ready, click Confirm and Create. You can also Cancel to return to Mapping or Validation without creating anything.

    Step 5: Results — Review what was created

    After the run completes, the Results screen shows:

    • Per-row outcomes — which rows succeeded and which failed, with a reason for any failures.
    • Summary counts — total Products and Items created.
    • Direct links to the newly created Products and Items so you can review them in Connect.
    • Next steps, such as configuring pricing, listings, or fulfillment for your new catalog entries.

    You can download the full run report for your records at any time. Every run — successful or not — is saved to Run History, where you can revisit the mapping, validation results, and outcome later, or use a past run as a reference when onboarding an updated catalog.

    Information

    After the Onboarding Assistant finishes running and accuracy is verified, the product enters draft status—ready to be published, versioned, and listed.

    Is this page helpful?
    Translate with Google
    Copied to clipboard