About PaymentLeaks
Ignacio De Navascués
Founder, PaymentLeaks
Over 12 years in payments infrastructure, correspondent banking, and international settlement. Has worked in the systems that move money between countries — SWIFT, ISO 20022, clearing, and instant payment networks. Built PaymentLeaks because no intelligence source spoke the language of the people who actually operate the infrastructure.
What is PaymentLeaks
AI-powered daily payments intelligence covering 300+ sources in 20+ languages. Every morning we produce a War Room with systemic impact analysis, regulation, competitive moves, parallel payment systems, and weak signals no other media covers.
Why it exists
“The payments industry moves too fast to rely on English-only sources.”
Editorial methodology
Every briefing is built on three verification pillars: (1) daily automated scan of 300+ sources — central banks and regulators (BCB, Fed, BOJ, RBI, FCA, BIS, FSB), trade press in Spanish, English, Portuguese, German, French, Italian, Mandarin, Korean, Japanese, Arabic, Hindi, Turkish and Indonesian, industry podcasts transcribed with Whisper, technical communities (Hacker News, Mastodon, GitHub releases), and operator newsletters; (2) semantic retrieval with pgvector and Qwen3 embeddings to link weak signals to historic archive events; (3) final headless Playwright verification of every cited URL — publication date, canonical domain, snippet present in HTML — before it enters the document.
No URL ships in a briefing without passing the verifier in the current session. No figure ships without backing in the cited source. No headline from the last 96 hours is recycled as if new. These rules are codified in scripts/briefing/verify_all.sh and block the commit if they fail.
Editorial standards
PaymentLeaks does not write headlines for clicks. The daily TOP 3 is picked by systemic impact — what move actually shifts the needle for someone running infrastructure — not by search volume. Predictions carry a horizon, an observable metric and a verifiable source; the ones that miss publish as MISS in the next briefing's Prediction Scorecard.
Anti-fabrication policy: zero tolerance for invented URLs, inferred figures, or quotes rephrased to dodge a checker. If a claim cannot be verified at its original source, it does not ship. If a section flags entity repetition, the item is replaced wholesale — never reworded to hide the name.
Corrections policy
If you spot a factual error, a misattributed quote, stale data, or a dead URL, email us at hola@paymentleaks.com. Material corrections publish in the next briefing under a "Correction" tag and the original stays on the historical record — past pieces are never silently edited.