Blog

Notes on regulatory compliance, document intelligence and how we build software in Mexico.

Cumplimiento

AML notices in XML: the most common format errors and how to prevent them with automated validation

A notice that is correct in substance can still fail to be accepted because of a format detail: a postal code missing its leading zero, a badly written amount or a code outside the catalog. We review the most frequent errors and how to catch them before filing.

Read article →
Ciberseguridad

The minimum security baseline for platforms that store client files

A practical checklist for those who build or operate systems holding IDs, tax certificates and proof of address: per-client isolation, encryption, API keys, signed webhooks, audit trail, retention and data handling with vendors.

Read article →
Ingeniería

MCP: how to expose a regulatory API to AI agents in a controlled way

The Model Context Protocol lets an AI agent use external tools. We explain why a sanctions-list check is a good tool for an agent and which controls must come with it so that a person still makes the decision.

Read article →
Ingeniería

Long, reliable processes: why we orchestrate document flows with a workflow engine

A document goes through OCR, classification, extraction and review, and sometimes waits days for a person. We explain why we moved away from hand-built queues and retries and use a workflow engine to orchestrate these processes.

Read article →
Cumplimiento

One year since the LFPIORPI reform: what changed for those who carry out vulnerable activities

On July 16, 2025, the reform to Mexico's Federal Law for the Prevention and Identification of Operations with Resources of Illicit Origin was published in the DOF. With the text of the law in hand, we review the changes that most affect the daily operation of obligated parties.

Read article →
Cumplimiento

Beneficial owner: how to identify and document it in an AML file

What the amended LFPIORPI says about the beneficial owner (beneficiario controlador), which documents usually show ownership and control of a legal entity, and how a well-organized file makes it easier to support that identification.

Read article →
Inteligencia documental

Reading Mexican documents with AI: why OCR alone is not enough

Voter ID (INE), tax status certificates, articles of incorporation and proof of address: what makes them hard to read well and why a KYC file requires more than turning images into text.

Read article →
Cumplimiento

SAT lists (articles 69, 69-B and 69-B Bis of the Federal Tax Code) explained for compliance teams

What the SAT publishes in each list, why they do not mean the same thing, and how to query them from a system, alongside OFAC and the UN, instead of checking them by hand.

Read article →
Inteligencia documental

Traceability in document extraction: every data point should point to the page and line it came from

In a regulated review, "the AI said so" is not an answer. We explain why every extracted data point must carry its source and how that changes audits, human review and the handling of discrepancies.

Read article →
Ciberseguridad

Vibe Coding: Why AI-Generated Apps Shouldn't Reach Production in Serious Businesses

Vibe coding lets you create functional apps in minutes, but what seems like an advantage hides critical risks in security, maintainability, and reliability that no serious business should take on.

Read article →
Tecnología

Not Ready for the Technological Leap: The Myth That Prevents Transformation

Many industries claim they're not ready for a major technological leap. The reality is they just need the right tool. Cursor's case demonstrates how a well-designed tool can transform a profession overnight.

Read article →
Tecnología

Cloud Provider Dependency: Why Small Teams Should Accept the Risk

AWS's recent outage that disrupted global services reminds us of cloud dependency risks. But is building your own infrastructure really better? We analyze why accepting this dependency is the right strategy for small teams.

Read article →
Empresa

Who we are

LoonByte builds software for regulated industries in Mexico: regulatory technology and document intelligence for anti-money laundering and identity verification.

Read article →