About us

MARPT is a rules-first Colombian immigration decision system — not a chatbot. It turns your answers into ranked visa matches, clear requirements, and risk flags, grounded in official sources (with Spanish as the legal source of truth).

The problem with generic AI assistants

General-purpose assistants summarize, draft, and rephrase well. They are optimized to generate coherent language. Immigration work is different: it requires jurisdiction-specific constraints, official sources, procedural steps, and the ability to surface discretion and risk.

A helpful paragraph is not the same as a case-ready answer. Applicants and law firms need a system that can: map facts to visa categories, flag missing or risky elements, and produce structured checklists aligned to the real process.

What MARPT is

Jurisdiction-aware immigration intelligence.

Official source anchoring. Built to reference and stay consistent with official government sources and process requirements.
Rule and eligibility engine. Models category fit using structured logic rather than free-form prompting.
Warnings and risk triage. Surfaces common rejection points, enforcement patterns, and procedural pitfalls when relevant.
Document checklist logic. Translates requirements into actionable lists tied to the selected pathway.
Bilingual by design. Spanish is the source of truth. English is an interpretation layer for clarity and review.
Structured outputs. Decision outputs, checklists, and risk flags—designed for reuse in forms, client intake, and review.
Explainable reasoning. Shows why a pathway is suggested and what facts materially change outcomes.
What MARPT is not
×
Not a chatbot. It does not “talk around” a problem; it produces structured, reviewable outputs.
×
Not crowdsourced guesses. It is not built on anecdotes as authority; sources and process constraints matter.
×
Not “prompt-only” advice. It is not dependent on clever prompting. The logic is encoded and maintained.
×
Not a replacement for a lawyer. It supports decision-making and case readiness. Legal judgment remains with the professional.
Assistants vs decision systems
AI assistants
marpt
Output type
Well-written text, summaries, drafts
Decision outputs
Structured matches, checklists, warnings, explainers
Source constraints
May mix sources; often not source-bound
Anchored
Designed to stay consistent with official sources and defined rules
Jurisdiction awareness
General by default; depends on prompts
Jurisdiction-aware
Built specifically for Colombian immigration categories and process
Discretion / risk
Often under-modeled or abstracted
Risk triage
Flags practical risk points, procedural pitfalls, and enforcement patterns when relevant
Structured checklists
Usually ad-hoc lists generated from text
Checklist logic
Requirement lists tied to visa pathways and conditions
Explainability
May be persuasive without being grounded
Explainable
Shows the inputs and constraints that drive a recommendation
Auditability
Hard to verify what changed across answers
Auditable structure
Rules, sources, and outputs can be reviewed and updated systematically
Who it’s for

Individuals

For applicants who want a clear view of options, requirements, and realistic risks—without relying on vague summaries.

  • See likely visa pathways based on your facts
  • Get structured document checklists
  • Understand common risk flags before filing
  • Export outputs for your records or attorney review

Law Firms

For firms that want consistent intake triage and clearer case readiness—while keeping professional judgment in the loop.

  • Intake triage and category fit consistency
  • Documentation readiness checklists
  • risk flags to guide follow-up questions
  • Client-facing explainers to reduce back-and-forth
  • Internal QA support for common edge cases
How we think about AI

AI is a tool inside MARPT, not the product. Language models help explain outputs and improve clarity, but they operate inside constraints: official sources, structured logic, and defined decision outputs.

The structure is the product. The value comes from jurisdiction-aware rules, bilingual source control, and outputs designed for real immigration workflows—not from generating paragraphs on demand.

MARPT doesn’t answer immigration questions. It models immigration decisions.

Use the Advisor to evaluate pathways and produce structured outputs you can review, compare, and share with a professional when needed.