Stripe's Adaptive Acceptance AI upgrade recovers $6 billion in falsely declined transactions in 2024
False declines cost US online retailers an estimated $81 billion in lost sales in 2023, affecting more than half of US customers and considered a major issue by 43% of retailers — with revenue losses from false declines often exceeding those from actual fraud.
How it works
Common implementation structure
How this type of workflow is generally built, generalized across documented cases — not tied to any one vendor's stack. Click any stage to read what happens there. Specific products that implement these stages appear in “Tools commonly seen” below.
Stage 1 · Issuer rejects legitimate payment
A legitimate payment is mistakenly rejected by an issuer as suspected fraud.
Tools used
Adaptive AcceptanceXGBoostTabTransformer+
Outcome
In 2024, Adaptive Acceptance recovered $6 billion in falsely declined transactions — a 60% year-over-year increase in retry success rate — while achieving 70% greater precision and reducing retry attempts by 35%. Model training time was cut from days to hours, enabling multiple deployments per week.
What failed first
Stripe's previous gradient-boosted tree model (XGBoost) saved billions of dollars in revenue but was limited in modeling the complex interactions among hundreds of factors influencing transaction success, and recent AI advances offered potential for improvement.