ITAD Pricing Intelligence Platform
A pricing intelligence platform that helps IT asset disposition teams ingest inventory, collect marketplace evidence, compare equivalent hardware, and review data-supported resale values.
A shared view of every moving part.
A multi-service workspace for turning large batches of used and refurbished IT equipment into reviewable pricing intelligence. Teams can upload or inspect product records, monitor scraping batches, compare marketplace listings, review product specifications and price history, apply rules, and inspect AI-assisted grades and price recommendations before deciding how assets should be valued.
Target users
ITAD operations managers, asset valuation teams, enterprise IT asset managers, refurbishers, and administrators responsible for pricing rules and market data quality.
Business problem
ITAD teams need to price many assets while marketplace data changes quickly and product descriptions are inconsistent. Manual research across channels makes it difficult to compare equivalent configurations, preserve a reliable price history, identify failed data collection, and apply the same business rules across a team.
Solution
We connected an inventory-facing Nuxt application with a Node.js operational API, a MongoDB service for raw scraped listings, and a Python AI service for scraping, matching, grading, formula generation, and price processing. Product batches move through collection, normalisation, specification matching, rule processing, and human review. Operators retain control through configurable rules, editable price details, audit-oriented statuses, and dashboard history.
Key features
Technology stack
Architecture
The Nuxt client communicates with a Node.js/Express service backed by Sequelize and PostgreSQL for users, products, rules, pricing, and operational records. A separate Node.js MongoDB service stores scraped product documents, scrape metadata, and failed collection details behind an HTTP API. The Flask-based Python AI service handles scraping helpers, product comparison, refurbished-grade detection, formula generation, price processing, and analytics using pandas, NumPy, scikit-learn, fuzzy matching, and Azure-hosted language models. Azure Queue and Storage integrations support asynchronous notifications, scraping jobs, and file-oriented workflows.
Timeline
Challenges
The platform had to reconcile two database models and several processing stages without losing traceability. Important challenges included preserving raw marketplace evidence while transforming records for relational workflows, matching equivalent hardware configurations, handling partial or failed scraping batches, making AI outputs explainable enough for review, and keeping rule versions and human overrides visible to operators.
Business outcomes
Platform views
Client feedback
“The platform gave our pricing team a much clearer path from raw inventory to a defensible resale value. Automated processing handles the repetitive work, while the review screens make it easy to understand and adjust the final decision.”