Skip to main content Skip to search Skip to main navigation
Diese Seite ist auch auf Deutsch verfügbar.
FULLTime eCommerce

AI Return Prediction & Analysis

Description

Highlights

  • Automatic returns analysis via scheduled task
  • Risk score (green / yellow / red) per product
  • Forecast via weighted moving average
  • Dashboard, rankings & reports inside the administration
  • Optional AI classification of return reasons (Anthropic Claude / OpenAI)

Features

  • Automatic daily returns analysis via scheduled task
  • Risk score (green/yellow/red) per product with configurable thresholds
  • Return forecast for 7, 14 and 30 days (weighted moving average)
  • Dashboard with KPI cards, time series and forecast box
  • Product list with sorting, filtering and risk ranking
  • Product detail view with return reasons and history chart
  • Return reasons categorized into seven classes
  • Optional AI classification of free-text reasons (Anthropic Claude / OpenAI)
  • Customer risk profiling with configurable thresholds
  • Direct jump to the customer profile in the admin
  • Automated email reports (weekly/monthly) with top-5 risk products
  • Instant test dispatch of the report
  • CSV export for product and customer lists
  • ACL roles: Viewer, Exporter, Settings
  • Shopware 6.7 native, PHP 8.2+, no data sharing with third parties

About the Extension

Returns Prognosis – Intelligent Return Analysis for Shopware 6

Returns are one of the biggest cost drivers in e-commerce. With Returns Prognosis, Shopware merchants gain full transparency over their return rates — per product, per customer, and with a look into the future.

What the Plugin Does

The plugin reads your order data daily (read-only access, no modification of shop data) and automatically calculates return rates per product. Products are classified into three risk levels: Green (normal), Yellow (elevated risk), and Red (critical). Thresholds are fully configurable.

Forecast Engine

Using a weighted moving average algorithm, the plugin calculates forecasts for the next 7, 14, and 30 days — with no cloud dependency, running entirely locally. For shops with the optional AI module enabled, free-text return reasons can be automatically classified via Anthropic Claude or OpenAI (own API key required).

Understanding Return Reasons

Every return reason is assigned to one of seven categories: Size/Fit, Quality, Product Description, Damaged, Wrong Item, Other, or Unknown. This lets you instantly see where your biggest optimization opportunities lie.

Customer Risk Profiling

The integrated customer list shows which customers generate an above-average number of returns. Configurable thresholds (minimum orders, return rate) prevent false positives. A single click takes you directly to the customer profile in the Shopware Admin.

Automated Email Reports

Receive weekly and/or monthly reports featuring the Top-5 risk products by email. Send time and recipient address are freely configurable. A test function sends the report immediately on demand.

Privacy & Data Security

The plugin exclusively reads existing Shopware order data — it never writes to Shopware core tables. No personal data is transmitted to AI services; only anonymized product texts (max. 200 characters) are used for classification. On uninstall, all custom tables and configuration entries are completely removed.

Technical Details

  • Compatible with Shopware 6.7
  • PHP 8.2+
  • Daily automatic recalculation via scheduled task
  • Full REST API for all functions
  • ACL roles: Viewer, Exporter, Settings
  • CSV export for product and customer lists

Details

  • Available: English, German
  • Latest update: 8 July 2026
  • Version: 1.0.0
  • Category: Administration

About the Extension Partner

FULLTime eCommerce

FULLTime eCommerce

Partner Status

  • Shopware Extension Partner Shopware
    Extension Partner

Details

  • Ø-Rating: 4

    Average rating of 4 out of 5 stars

  • Partner since: 2014
  • Extensions: 8

Support

  • Based in: Germany
  • Speaks: German, English
  • Response time: Quickly