Skip to main content
NewVantaSoft Agent Service: custom AI agents for your business
VantaSoft
All projects
Goodwill Michiana
Nonprofit Retail

E-Com Eye: E-Commerce Intake for Donated Goods

An in-store system that finds donated items worth selling online, routes them to Goodwill's e-commerce hub, and tracks each one from the pricing table to the sale.

Client
Goodwill Michiana
Industry
Nonprofit Retail
Timeline
2026
Platforms
Web, Tablet

Who uses it

  • Store pricers
  • Processors
  • Hub sorters
  • E-commerce managers

The brief

What Goodwill Michiana needed

Goodwill Michiana sells donated goods in its stores and online through ShopGoodwill. It wanted to grow its e-commerce revenue, and that meant getting better at three things: picking the donated items worth selling online, closing the feedback loop between its stores and its e-commerce hub, and tracking every item from the person who spotted it to the customer who bought it.

The people doing the picking are store pricers and processors, not e-commerce specialists. The system had to fit into their normal work and help them get better at it over time.

We built E-Com Eye with Thriftly, the retail technology company whose pricing system Goodwill Michiana uses.

What we built

What we built for Goodwill Michiana

  • Flagging in the pricing app

    During normal pricing, items from high-value brands are flagged for e-commerce automatically. Pricers can flag anything else with one tap, and each flag prints a barcode label that stays with the item.

  • Processor tablet app

    Processors scan a flagged item, take four photos (front, back, brand tag and details), grade its condition on a six-step scale, and send it to the hub or price it for the floor.

  • AI listing drafts

    A vision model drafts the listing title and description from the processor's photos, so the hub starts from a written listing instead of a blank form.

  • Sorter review app

    Sorters at the e-commerce hub see the photos and the grade, then accept, reject or escalate each item, with a reason every time. Items reach the sorter who specializes in their category.

  • Manifest and floor pricing

    Accepted items are added to the hub manifest automatically. Rejected items come back with a recommended price and print a tag on the right printer for clothing or hard goods.

  • Manager dashboard

    Pricer yield, processor throughput, acceptance rates by store and by person, and the full history of any single item.

How it works

Step by step

  1. 1

    Flag

    A pricer flags the item, or the pricing app flags a high-value brand on its own. A barcode label is printed.

  2. 2

    Photograph and grade

    A processor scans the label, takes four photos on a tablet and grades the condition.

  3. 3

    Review

    A sorter at the hub accepts, rejects or escalates the item and says why.

  4. 4

    List or return

    Accepted items ship to the hub on the next manifest. Rejected items get a price and go back on the sales floor.

Results

What it changed

  • Every item tracked from store to sale

    Managers can see who flagged an item, who graded it, which store it came from and whether it sold, along with yield and acceptance rates by store and by person.

  • Store staff learn from every decision

    Each accept or reject reaches the pricer and processor with the reason, so the people choosing items for e-commerce get better over time.

  • Rejected items stay in the store

    An item the hub turns down gets a recommended price and goes back on the sales floor. It never leaves the store.

Where it stands Built between May and September 2026 and handed to Goodwill Michiana's team, which is planning the rollout to more stores.

Stack

  • React
  • TypeScript
  • Go
  • Vision LLM
  • Tailwind CSS

What we did

  • Custom Software
  • Web Applications
  • LLM Integration
Next project Amira Learning Teacher Reporting Portal for AI Tutoring Platform
View project

You run the business.We run the tech.

A long-term partner, not a one-off project. The IP is always yours. It starts with a conversation.