case study

Leveraging AI to Turn Tanger’s Loyalty Data into Dollars

KUNGFU.AI helped Tanger Outlets transform 14 million messy receipt images into actionable insights with a custom AI solution, achieving over 90% accuracy and unlocking a new revenue stream.

AI Solution(s)
Computer Vision
Industry
Retail
Tanger Outlets Case Study: Receipt Data Extraction

Receipts, Rewards, and Revenue

Vision

Monetize Millions of Receipts

Tanger Outlets, a leader in outlet shopping with 38 centers across North America, had a goldmine of untapped data—14 million customer receipt images submitted through its loyalty program. The goal? Extract valuable information from these receipts and turn it into actionable insights. But with no standardized format, manually entering the data was both tedious and cost-prohibitive. Tanger knew AI was the answer and called KUNGFU.AI to help.

Challenge

Crumpled, Crooked, and Chaotic Data

No two receipts were the same. Vendors used different formats, lengths, and layouts. Some receipts were poorly lit, others were crumpled or incomplete. Extracting structured data like tax amounts, item descriptions, discounts, and SKUs from such an inconsistent dataset was a daunting task. Tanger needed an AI solution that could tackle this chaos while ensuring accuracy and speed.

Breakthrough

Feasibility Sprint to Full Production

KUNGFU.AI proposed a six-week feasibility sprint to test various computer vision techniques and determine project viability. Competing against another vendor in a proof-of-concept bake-off, KUNGFU.AI earned Tanger’s trust and proceeded with a 16-week production phase, followed by an eight-month improvement and maintenance phase.

Key Highlights:

  • Feasibility Sprint: Explored models like RetinaNet, YOLO v3, and Faster R-CNN to identify the best approach.
  • Production: Developed a custom model with sub-3-second latency, achieving over 90% accuracy across 100+ receipt variants.
  • Maintenance & Improvement: Expanded receipt coverage, boosted accuracy, and optimized processing speed.

The final solution was deployed on Microsoft Azure, using a Dockerized microservice with RESTful architecture that accepted JPEG images and returned extracted data in a clean JSON payload.

Outcome

New Revenue Stream, Minimal Effort

KUNGFU.AI delivered a cutting-edge receipt-scanning application that exceeded expectations. The AI model now extracts information across 12 fields, including city, state, tax, SKUs, discounts, and item descriptions—all in under three seconds. Tanger gained a scalable tool that saves thousands of hours in manual labor and unlocked an entirely new revenue stream.

Project Highlights:

  • +90% Accuracy
  • 12 Extracted Data Fields
  • 100+ Receipt Variants Supported
  • ≤ 3-Second Latency

Tanger Outlets is now equipped with cutting-edge AI technology that turns raw data into powerful insights, reinforcing its leadership in the retail space and delivering significant operational savings.

Download the Case Study

quotes icon

See how KUNGFU.AI’s strategy-led engineering bridges strategy and execution for AI success.

Work with Us

More case studies