{"version":1,"type":"story","url":"https://digestai.news/story/galgo-cuts-fraud-evidence-to-under-2-using-zapier-s-gpt-5-mini","json":"https://digestai.news/story/galgo-cuts-fraud-evidence-to-under-2-using-zapier-s-gpt-5-mini.json","markdown":"https://digestai.news/story/galgo-cuts-fraud-evidence-to-under-2-using-zapier-s-gpt-5-mini.md","slug":"galgo-cuts-fraud-evidence-to-under-2-using-zapier-s-gpt-5-mini","headline":"Galgo cuts fraud evidence to under 2% using Zapier's GPT-5-mini","summary":"Ignacio Piñeiro, operations lead at Galgo, a Mexican fintech that finances motorcycles, built an AI‑driven fraud check after noticing that manual photo reviews took about three minutes each and cost roughly 150 hours per month. The company receives more than 3,000 delivery photos monthly, each required to prove a bike was actually delivered to the borrower. Using Zapier, the photos are sent to OpenAI’s GPT‑5‑mini, which validates whether a real customer and an eligible vehicle appear together. The model returns a VALID or INVALID verdict in seconds, and a second model tags the reason for any rejection. Only the flagged photos are sent to human reviewers via Google Sheets and Slack.\n\nSince the system launched in January 2026, Galgo reports that invalid delivery evidence dropped from roughly 8% to under 2% of submissions. Manual review time fell from 150 hours per month to fewer than 60 photos needing human attention. Piñeiro estimates the reduction prevents tens of thousands of dollars in first‑payment defaults and frees the fraud team to focus on truly suspicious cases.","keyPoints":["Invalid delivery evidence fell from about 8% to under 2% after AI deployment","Manual review time dropped from 150 hours per month to under 60 photos reviewed","System automatically processes over 3,000 photos each month using GPT‑5‑mini via Zapier"],"whyItMatters":"The case shows how exception‑based AI can slash manual labor and fraud losses in fintech, offering a template for other industries to automate high‑volume verification tasks.","category":{"slug":"enterprise","name":"Enterprise & Industry","url":"https://digestai.news/category/enterprise"},"entities":{"companies":["Galgo","Zapier","HubSpot","Google","Slack"],"models":["GPT-5-mini"],"people":["Ignacio Piñeiro"]},"firstPublishedAt":"2026-09-23T00:00:00Z","updatedAt":"2026-09-23T00:00:00Z","sourceCount":1,"hasPrimarySource":false,"sources":[{"outlet":"Zapier Blog","title":"How Ignacio Piñeiro scaled fraud control with exception-based AI review","url":"https://zapier.com/blog/galgo-delivery-evidence-ai-review","publishedAt":"2026-09-23T00:00:00Z","type":"press","primary":false,"lead":true}],"sourceNotes":null,"discussions":[],"thread":null,"cite":{"text":"Digest AI, \"Galgo cuts fraud evidence to under 2% using Zapier's GPT-5-mini\", 23 September 2026, https://digestai.news/story/galgo-cuts-fraud-evidence-to-under-2-using-zapier-s-gpt-5-mini","publisher":"Digest AI","title":"Galgo cuts fraud evidence to under 2% using Zapier's GPT-5-mini","datePublished":"2026-09-23T00:00:00Z","url":"https://digestai.news/story/galgo-cuts-fraud-evidence-to-under-2-using-zapier-s-gpt-5-mini"},"generatedBy":"Written by Digest AI's editorial model from the linked sources; the sources are the record.","license":"Headlines, digests and key points are written by Digest AI and may be quoted with a link to the story page. Linked articles belong to their publishers. Terms: https://digestai.news/terms#reuse"}