New Product Introduction

Automate translations and approval process for a new product.

Automate the approval and onboarding process for the newly created products. The template helps you build a process that changes the traditionally manual process by embedding AI-driven translations, real-time notifications, and dynamic updates, ensuring faster time-to-market and consistent compliance. This template is designed to show the possibilities and capabilities, you can use it as a starting point and customize it according to your specific business needs.

By decoupling business logic from complex code, this process empowers you to launch new products, iterate, and scale more efficiently, while maintaining accuracy and transparency across teams.

Key features

Feature / Component
Description

AI translation

Automatically translates product information into multiple languages using AI.

Low-code automation

Automates and configures approval processes without requiring deep technical expertise.

System connectivity

Approval, translation, and update steps are integrated into a unified process across systems.

Faster innovation

Teams can iterate and launch products quickly without bottlenecks from manual coordination.

Faster iime-to-market

Moves products from creation to approval in a short time.

Improved accuracy

Reduces human error in product data handling and translation.

Transparency & compliance

Real-time tracking of approval decisions ensures clear auditability and compliance.

Process template flow

The Emporix product introduction process makes sure every new product is handled in a clear and consistent way. When a product is created, its details are retrieved, the system fetches relevant user groups and automatically assigns the CE Manager and CE Admin roles. The AI then generates descriptions and translations for all required languages and shares them with the assigned approvers by email. Approvers receive a form containing the localized product information and can either approve, which publishes the product along with its translations, or decline with suggestions. If declined, the approver provides feedback and suggestions - the suggestion field is only visible when "rejected" is selected in the form. Then, an email is sent to the requester, who manually updates the translation based on the feedback. Once updated, the approver receives a new approval request. If the approver declines a second time, the process stops and the requester receives a notification that the translation was not accepted.

The value stream begins with a trigger and includes several subsequent steps.

Process trigger

The process starts when a product is created or updated. It listens to the product creation/update events configured in your commerce and automatically initiates as soon as such an event occurs. Once triggered, the process retrieves all relevant product details, including metadata such as the product name, attributes, specifications, brand, categories, and any initial content provided by the creator. The system verifies that the product exists and confirms that it is ready for further processing, such as translation and approval.

Process steps

Product translation

After the product details have been retrieved, the process triggers an AI Agent responsible for generating the required translations of the product descriptions. Using the product’s content as the source, the AI automatically produces localized descriptions. This step generates the text that is later reviewed by the approver. The translated content is stored as a draft and becomes the basis for the subsequent approval process.

By default, the Translation Agent is configured to translate product information from English to German. However, you can adjust the agent's prompt to translate to different target languages. For example:

Name and description are the localized fields in product data.

The AI used for the process is the Translation Agent designed by Emporix. To learn more about the way the agent works, see the Translation Agent documentation.

Users assignment

After the product has been created and its details retrieved, the process proceeds to identify the users responsible for managing and approving the product. The system queries the user directory to retrieve all users belonging to the CE Manager group, who act as the requestors in this process and are responsible for initiating and overseeing the product introduction. It then retrieves users belonging to the CE Admin group, who serve as approvers and who review and validate the generated translations. Once these users are identified, the process automatically assigns the appropriate manager and approver roles, ensuring that responsibility and approval paths are clearly defined before moving to content generation and review.

Email requesting translation approval

Once the AI-generated translations are ready and the relevant CE Manager and CE Admin have been identified, the process prepares and sends an approval email to the assigned approver.

The email includes a structured approval form that displays the requester’s name and email (automatically fetched from the previous steps), followed by a table listing the products awaiting approval. For each product, the form presents key information such as the Product ID, Product Name, and the translated description. It also provides fields where the approver can propose changes to the product name or description if needed. The form requires the approver to explicitly indicate whether each product should be accepted or rejected. This form allows the approver to review the translated content in a consistent format and submit their approval or feedback directly from the email. The approver can either approve the translation which results in the product update, or requests corrections.

Rejecting the translation and change requests

If the approver rejects the translation during the review, they can provide suggestions and comments. The proposal field and comment fields only appear when "rejected" is selected in the approval form. In the input fields the approver can specify the required changes to the product name and description. By submitting the form, the approver communicates exactly what needs to be improved, enabling the process to pass this feedback into the next step, where the requester can update the translations based on the reviewer’s notes.

Translation corrections

After the requester receives the rejection email with the approver's feedback, they manually update the translation based on the suggestions provided. Now, the requester directly edits the product descriptions and translations, updating the text to reflect the approver’s revision instructions. Then, the process automatically sends a new approval email to the approver, containing the updated translation for review.

Process completion

When the translation is correct, the approver accepts it. Upon approval, the product is updated accordingly, marking it as localized, and a success email is sent to the requester confirming that the translation has been accepted and the product has been published.

Second rejection scenario

If the approver declines the translation a second time after the requester has made manual corrections, the process stops and does not loop further. In this case, the requester receives an email notification that the translation was not accepted, and the process ends without translating the product.

  • To learn more about the AI agents in Emporix, see Agentic AI.

  • To learn more about the way forms are processed in the flow, see Forms.

Last updated

Was this helpful?