National Platform for Algorithmic Governance · Guatemala
The next form of the state is not digital government. It is governed intelligence.
NaciluzIA uses AI to help the State serve citizens in minutes instead of weeks, in their own language — while a human official, never an algorithm, makes and signs every binding decision.
Cada decisión, a la luz.
Our mission
To let the Guatemalan state serve every citizen in minutes instead of weeks, in their own language, while keeping every binding decision in human hands — and to prove that government technology can be clean, fast, transparent, and inclusive, then share that model across Central America.
What we will not do: never let AI make a binding decision · never claim to “catch” corruption, only surface evidence for authorized human review · never lock the government into a single vendor.
The problem
The State can’t use the information it already holds.
Siloed systems and manual processes mean leakage, delay, and exclusion — and citizens who speak a Mayan language are shut out of a Spanish-only digital state. In Latin America a single government transaction takes 5.4 hours on average, and only 7% of people ever finish one fully online (IDB, 2018).
Leakage
Ghost and duplicate beneficiaries drain social programs; discovery is manual and after the fact. Latin America wastes 4.4% of GDP (~$220B/yr) in public spending (IDB, 2018).
Delay
Citizens travel hours and wait days — often to be turned away for a missing document — across 340 municipalities.
Exclusion
25 recognised languages and 43.6% of the population indigenous (INE, 2018) — yet public digital services are built almost entirely in Spanish.
How it works
A system where AI advises and a human decides.
Each layer has one job and hands a clean, verified result to the next. Between any AI suggestion and any real-world effect sits a gate that cannot act without both a deterministic rule result and a human signature.
Audit layer — every action hashed, timestamped, and signed under Decreto 47-2008.
The product
A citizen applies on WhatsApp, in their own language.
Instead of a four-hour trip and a two-week wait, a citizen checks eligibility and gets a correctly-drafted application in minutes — then a human official reviews and signs it.
- Multilingual intake — Spanish and Q’eqchi’ to start, voice-first.
- A three-gate eligibility check: valid DPI, no duplicate benefits, meets the threshold.
- An application drafted automatically and routed to an official for signature.
- Every step written to a tamper-evident ledger.
Citizen message
Nin kʼo wi ri nu DPI, xbʼe utzalaj?
NaciluzIA
Usalajixik. Niqʼat ri aw DPI rikʼin RENAP...
Result
Yalan utzalaj. Taq'alajisab'al ri taq na'oj richin ri ajchajinel.
By the numbers
The stakes, and the opening.
wasted each year in LatAm public spending
IDB, 2018
of every public contract lost to corruption
UNODC
AI-in-government market, 2024→2033
Grand View, 2024/25
LatAm countries with any national AI strategy
OECD
Transparency & anti-corruption
Discovery becomes automatic — and defensible.
Continuous cross-institutional checks surface ghost beneficiaries and conflicts of interest as evidence for human investigators — the platform does not determine misconduct. Every decision is hashed into a tamper-evident chain. “AI makes corruption automatically discoverable, and leaves judgment to people” is a claim you can defend.
Cross-check data
Surface anomalies
Human investigation
Accountability runs both ways
The citizen’s right to see and appeal.
A decision no one can question is not accountable. The same transparency that lets an auditor read a decision lets a citizen challenge it.
See the reason
Every decision returns the plain-language rule that produced it and the data it used — in the citizen’s language.
Appeal to a person
Disagree, and human review opens a signature-tracked task with a deadline and escalation.
Trace the record
Because every step is in the ledger, a reviewer or the Contraloría can reconstruct exactly what happened.
Correct & reverse
A decision made under a rule later found wrong is traceable and reversible — and the citizen is told.
Credibility through candor
What happens when the AI is wrong.
The governing property: no failure can produce a binding action on its own. Whatever goes wrong, the system degrades to human handling — never to a guess.
A wrong recommendation
Logged, never executed; a human sees the proposal, its sources, and the rule result before signing.
The AI is unavailable
The request routes to the ordinary human queue; service continues without AI.
A rule is outdated
Rules are published and versioned; corrected, and every decision under it is traceable and reversible.
The ledger is tampered with
Recomputing the hash chain detects the alteration immediately.
Why institutions cooperate
A reason for every partner to say yes.
The hard part is coordination, not technology. NaciluzIA strengthens each institution's own mandate — it does not police officials.
Faster service, cleaner reporting
Faster service, fewer duplicate and ghost payments, cleaner donor reporting — approval authority unchanged.
Controlled identity checks
Controlled, least-privilege identity checks with a record of every query — not bulk data sharing.
A complete evidence trail
A complete, tamper-evident evidence trail that strengthens the audits it already runs.
A visible digital win
A visible, exportable digital-transformation win aligned to the national plan.
Why now
This is what unaccountable government costs.
In 2015, Guatemala's “La Línea” customs-fraud ring skimmed roughly $328,000 every week; when it surfaced it brought down the president and vice-president (InsightCrime, 2015). The lesson is not that people are corrupt — it is that discretion without a traceable record is the vulnerability. For the first time, there is a tool that closes it without seizing power from institutions. Rising demand for responsible AI, active donor interest, and a region with almost no rulebook yet make now the moment to set the standard.
Open standard, not a vendor
Public infrastructure, built in the open.
NaciluzIA is designed as an open, auditable governance standard aligned with the IDB Group's AI Framework — stewarded by a nonprofit foundation and governed by a multi-stakeholder board (universities, indigenous-language representatives, legal, cybersecurity, and civil society), with the State keeping authority, ownership of every decision, and the signature. Built with Guatemalan universities.
Legally defensible
Deterministic, human-signed decisions under the Ley de Firmas Electrónicas (Decreto 47-2008).
Inclusive by design
Voice-first, multilingual, built for low-literacy and offline users from day one.
Fundable
Aligned to the IDB / CABEI responsible-AI agenda and the GAE’s 2025–2030 digital plan.
Insights
Ideas on governed intelligence.
Essays on accountable AI in government, transparency, and serving citizens in their own language.
Introducing NaciluzIA: Light on the Public Sector
An older woman in Alta Verapaz who speaks only Q'eqchi' is owed a pension. To claim it, she travels four hours to a municipal office, waits most of a day, and is…
Read →AnchorsWhy a Government AI Must Never Decide
There is a seductive pitch making its way through ministries across Latin America: let artificial intelligence approve the benefits, issue the permits, flag the…
Read →VisionGoverned Intelligence: The Next Form of the State
For two decades, the ambition of public-sector modernization had a name: digital government. Put the forms online. Move the records into databases. Let a citizen…
Read →VisionThe Limit Is the Unlock
Most conversations about artificial intelligence in government begin with a question about capability: how much can the machine do? We want to begin somewhere…
Read →VisionWhy the Moat Isn't the Software — It's the Standard
Ask what makes a government-technology company defensible and the instinct is to point at the software: the model, the interface, the pipeline that turns a…
Read →VisionA Public-Sector Operating System
We are often introduced as an AI company, and we understand why. Artificial intelligence is the part of NaciluzIA that catches the eye — the system that…
Read →The team
Built by researchers and forward-thinking innovators to transform Guatemala’s economic landscape.
Our team pairs deep macroeconomic understanding with the technological agility required to catalyze AI transformation across government, industry, and local universities.
Business leader with more than 20 years of experience in corporate affairs, sustainability, and strategic economic development across international public and private sectors. He is currently CEO of Innovative Strategies, an international firm specializing in political and economic risk analysis, investment advisory, and sustainability, helping organizations manage risks and capitalize on opportunities in complex markets.
Previously, Dr. Rosales served as Guatemala's Minister of Economy and Chief of Staff to the President, where he led international investments, trade agreements, and competitiveness initiatives while serving on the national monetary board and boards of banks and energy institutions. He has also held strategic leadership roles at Corporación Castillo Hermanos, the Guatemalan Chamber of Construction, and Grupo Pantaleón, and is an active member of Concordia Americas Summit, Fundesa, and the Americas Society Council of the Americas.
He holds two doctorates in economic sustainability and public administration and governance, as well as an MBA magna cum laude specializing in finance, communication, security, and sustainability. He is a frequent keynote speaker at global forums in Europe and Latin America and has wide experience as a senior consultant in investment and strategic advisory for multinational companies, private-sector chambers, and research institutions.
Let’s build the pilot together.
If you work in government, at a multilateral, at a university, or in civil society and this resonates, we'd like to talk.