konaai-newsletter-MIT Whitepaper

Detect Hidden Fraud with Proven Analytics

How a Fortune 100 company operating in 190 countries uncovered ~$75 million in potential cost recovery hidden in split purchase orders — and a look inside konaAI’s agentic AI platform for bribery, corruption, and risk detection. 

The $75 Million Split: Uncovering Price Variance Fraud in Purchase Orders

  • The Setup: A Fortune 100 company operating in 190 countries partnered with konaAI to process and review 24 months of SAP Source-to-Pay data across 400+ vendors.
  • The Scheme: PO requestors split purchase orders for the same material and plant across vendors at varying prices within 3-day windows, bypassing delegation-of-authority thresholds and vendor share-of-business limits.
  • The Result: ~$75M in potential cost recovery identified, with price variances of 25–45% in raw materials, ~14,000% in packaging materials, and ~92,000% in chemicals.
“The algorithms that we use can be run on any data for any company to detect fraud, risk, or corruption”
Related on the Website
 
 

Agentic AI that Looks Beyond Detecting Bribery, Corruption, and Fraud Risk

  • Coverage: Machine learning tests run across every P2P, T&E expenses, and O2C transaction, fully mapped to your ERP systems.
  • Agentic AI: Purpose-built agents now handle document classification and screening, contract compliance, PO alignment, segregation of duties, culture pulse and more, each reachable through a digital assistant that you can just ask.
Related on the Website
 
 
“We look at the breadth of compliance to see where AI and data analytics can be of real use to professionals in the industry”