E-Commerce

Categorizing and risk rating products

Improving product safety and compliance in e-commerce

Challenge

Understanding risk within e-commerce

E-commerce platforms, especially those using third-party sellers like Amazon, Shopify, and eBay, need to make sure their products are safe and meet regulations. Manually reviewing product titles and descriptions to find potential risks is slow and often leads to mistakes.

Solution

Efficient Risk Categorization with OmniAI

By leveraging AI language and classification models OmniAI transforms lengthy manual reviews into swift, accurate assessments. In this case study, we’ll do the following:

Data collection
OmniAI connects to the Amazon product listings, pulling titles and descriptions.
Risk identification
Custom AI models are trained to identify and categorize potential risks like choking hazards, fire risks, and battery malfunctions.
Data integration
The identified risks are then cataloged and integrated into the Omni warehouse for easy access and analysis.
Rapid onboarding
The time to onboard new sellers is significantly reduced, from over 48 hours to mere minutes, regardless of the SKU volume.

Get started in a few lines of code

Here, we’ve defined possible values for the hazard types (FALLING_HAZARD, BATTERY_MALFUNCTION, CHOKING, etc.) along with any additional fields we want to extract.When we run the sync, we’ll have a table that combines our existing data, with the newly extracted columns.

1product_hazards:
2  description: "Array of possible hazards associated with a product."
3  options:
4    - "FALLING_HAZARD"
5    - "BATTERY_MALFUNCTION"
6    - "CHOKING"
7    - ...
8  type: "String"
9
10recipient:
11  description: "Who this product was purchased for"
12  options:
13    - "SELF"
14    - "CHILDREN"
15    - "SPOUSE"
16    - "PARENT"
17    - "FRIEND"
18  type: "String"

Battle tested tools for top use cases

The company now has a dynamic database that automatically updates with risk assessments for each product. From here, we can query this table directly with SQL, or ingest it into a variety of BI tools.

Insights

From manual review to an automated system

Bringing in OmniAI sped up adding new sellers and also offered great insights into product safety. For instance, an unexpected rise in BATTERY_MALFUNCTION risks in a specific category of products was quickly identified and addressed.Switching from manual checks to an automated AI system, OmniAI has helped our client create a safer and more rule-following e-commerce world. This builds trust and reliability for everyone buying and selling.