FASE 01
Observe
Map real workflows: what enters, who decides, what repeats. No production changes.
Applied Intelligence Studio · Mendoza · Buenos Aires · United States
Autonomy with
We build agentic offices for companies that need to operate with AI without losing control.
We design, implement and govern agentic systems with multi-model capability: we evaluate 50+ local and cloud models, then choose the right combination for each process. Agents connected to CRM, WhatsApp, email, documents, data and real operations — with memory, permissions, logs, human review and rollback.
Mendoza base
Buenos Aires market
United States clients
LIVE AGENT SIM
Customer operation agent
Inbound request
Customer asks for status and options.
ERP/CRM lookup
The agent retrieves permitted data.
Runs procedure
Applies policy, drafts answer and task.
Governed and logged
Trace, risk score and audit log.
CID-9128Correlation ID
18/100Risk
v14Prompt
REAL OPERATIONS
Mendoza base. Presence in the United States and Buenos Aires. Expanding from Latin America.
The public site must show scale without inventing logos: operating presence, multi-model capability, methodology and governance. Sensitive clients stay unnamed until permission exists.
Regional and international presence
Mendoza base, presence in the United States and Buenos Aires, with a working network expanding from Latin America.
Serious agentic capability
We evaluate local and cloud models, define hardware, software and model load per operation, and govern every sensitive action.
Executive delivery
Free self-assessment, Executive Radiography, governed pilot, runbooks, follow-up and dashboards so owners see what is happening before scaling.
BRAND ENTRY
SimulaciónUNO is the public entry point.
The same studio is often searched as SimulacionUNO, Simulacion Uno, Simulacion 1 or S1. The canonical public site is simulacionuno.com.
SimulaciónUNOSimulacionUNOSimulación UNOSimulacion UNOSimulación UnoSimulacion Unosimulacionunosimulacion unoSimulación 1Simulacion 1
PUBLIC SIGNAL
Weekly technology column on Channel 7 Mendoza.
Every Sunday, Actividad Económica AI News covers agentic offices, applied AI and business use cases for a Mendoza business audience. Powered by SimulaciónUNO.
MULTI-MODEL CAPABILITY
Local models, cloud models and business systems under one governed layer.
We do not sell a model. We test the leading options, choose the right load for each process and define the hardware, software and human approval points for each operation.
Integration
n8n
Integration
Gmail
Integration
Airtable
Integration
HubSpot
Integration
Salesforce
Integration
Notion
Integration
CRM
Integration
ERP
Integration
AWS
Integration
50+ models evaluated
Policy-based selection by cost, latency, privacy, task type and risk.
Infrastructure per project
We define the hardware, software, environments, credentials and governance each operation needs.
Human approval gates
Sensitive actions stop, explain themselves and wait for the owner.
SECOND STEP · PAID PROFESSIONAL SERVICE
This is not the free self-assessment. It is an Executive Radiography of one real workflow.
The self-assessment provides an indicative self-reported 1–100 score. Radiography starts only if you want to move forward: over 5–10 business days we validate evidence, systems, risks, architecture and costs to decide what should be implemented.
Immediate orientation
Free self-assessment
Eight self-reported answers produce an indicative 1–100 score. It does not audit company documents, systems or results.
Professional analysis
Executive Radiography — USD 950
We study one real workflow through interviews, evidence, systems, architecture, risks, implementation options and an executive report.
USD 950
5–10 business days · one real workflow
The prior framing call is free, lasts up to 30 minutes and is not this Radiography.
HOW THE USD 950 CREDIT WORKS
If within 30 days you contract the Governed Pilot, Initial Build or a Corporate implementation for this same workflow, the full USD 950 is deducted from the first milestone. It does not apply to monthly operations or another workflow.
One selected business workflow
Up to 3 stakeholders
Process map + baseline
Systems, data and permissions involved
Private architecture + estimated costs
Risks, human approvals and rollback
Two paths: configure/buy or build
Decision: proceed, reformulate or stop
Implementation scope, budget and roadmap
Executive decision report
HOW WE WORK
First we understand your operation. Then we transform it.
Four phases. No big bang: your current office keeps running while the second office learns beside it.
FASE 02
Duplicate
Build the operational twin: shadow-mode agents recommend without writing or contacting.
FASE 03
Measure
Compare human vs. twin with hard metrics: speed, quality and decision alignment.
FASE 04
Integrate
Enable minimal reversible write access with traces, policy, human queue and rollback from day one.
DATA AND WORKFLOWS
Your systems record. They do not coordinate.
The problem is not choosing the newest AI model. The problem is that data, decisions and workflows are often trapped in ERP, CRM, spreadsheets, WhatsApp and key people. SimulaciónUNO builds a governed operating layer to free context, coordinate work and automate without breaking what already works.
Not AI on a locked box
A chatbot on top of isolated systems may sound good, but it will not transform operations if it cannot reach the right context.
A new operating layer
The agentic office coexists with ERP, CRM and existing tools, but demotes them: they become connected sources, not cages.
Govern first, automate second
Data, permissions, human review, logs, metrics and rollback are designed before sensitive actions are scaled.
OWN METHOD
AI-native applied to real companies.
SimulaciónUNO’s public thesis is direct: turn real processes into governed agentic offices, with measurement, evidence and human control from the first pilot.
Operational reading
We start from concrete processes: customers, repeated tasks, scattered data, follow-up, reports and decisions that depend on key people today.
Own method
SimulaciónUNO turns that map into pilots, dashboards, agents and governed operations without exposing internal sources or sensitive methodology.
Visible control
AI-native work enters with permissions, logs, human review, rollback and measurable evidence from the first pilot.
PROOF OF CONTROL
The governance layer is the product.
Method observe → duplicate → measure → integrate: we build an operational twin next to your operation, measure it against the real process and only integrate it after it proves control, improvement and traceability. All under Trust Core.
3D fallback activo: imagen procedural liviana para mobile/reduced-motion.
38%
less operating time in repetitive processes
99.9%
target SLA for enterprise plans
12+
modules to build AI Native Organizations
4
governance frameworks mapped in the dashboard
MODULES
Services built to operate, not decorate slides.
Composable modules to build an AI-native organization with real controls from the first pilot.
core
Agent Orchestrator
Multi-agent engine connected to real systems.
Automates end-to-end tasks with traceability and human control.
42% faster cycle time
Multi-LLM routingMCP readyDev/Test/ProdAudit logsRollback
growth
Agentic CRM
Sales, follow-up and pipeline with actionable AI.
Turns conversations into opportunities, tasks and reports.
+23% qualified pipeline
Lead scoringWhatsAppEmail
cx
Omnichannel CX
Agents for chat, email, voice, SMS and WhatsApp.
Resolves repetitive requests and escalates exceptions with context.
70% autonomous resolution target
ProceduresQA queueHITL
data
RAG Knowledge Hub
Secure knowledge base for citable answers.
Reduces hallucinations and keeps sources controlled.
92% grounded answers target
RetrievalCitationsPII guard
risk
Decision & Scoring
AI for risk, eligibility and prioritization.
Explains decisions, signals used and review thresholds.
30% faster review
ExplainabilityRisk tiersPolicy engine
growth
Intelligent Outreach
Prospecting, enrichment and audited campaigns.
Personalizes without losing brand, consent and trace control.
+18% reply rate target
WaterfallsDeliverabilityConsent
marketing
AI Web + GEO
Sites and content optimized for generative engines.
Improves discovery across search, generative assistants and answer engines.
Share-of-voice dashboard
Schemallm.txtFAQ blocks
data
LATAM Data Products
APIs and dashboards with macro and sector data.
Integrates BCRA, BCB, Banxico and vertical data sources into AI workflows.
Regional data advantage
APIsDashboardsFreshness
FLAGSHIP DASHBOARD
Trust Core
A governance dashboard proving that every agent has an owner, risk class, traceability, continuous evaluation and a compliance path.
GOVERNANCE CONTROL PLANE · DEMO SANDBOX
Live governance cockpit
Mock data, real product logic: every agent is mapped to owners, controls, risk classes, traces and review flows.
healthy
Agent Inventory
42 agents
ISO 42001 §6 · AI Registry
healthy
Continuous Risk Score
18/100 avg
NIST AI RMF · MEASURE
healthy
Traceability / Audit Log
100% traced
OWASP LLM05 · Observability
warning
Drift Detection
2 warnings
Faithfulness · Relevance · Safety
healthy
Cost & Tokens
USD 842 MTD
Cost tracking
warning
Human Review Queue
9 open
EU AI Act · Human oversight
danger
Policy Violations
3 blocked
OWASP LLM01/02/07
healthy
Model/Prompt Rollback
v14 stable
Incident management · Versioning
healthy
Compliance Map
78% mapped
NIST · ISO · OWASP · CSA AICM
warning
EU AI Act Readiness
Classified
Risk classification · Annex III
FIG. 01 · RISK / COST / EVALS
Continuous measurement
| Lun | 22 | 91 |
| Mar | 19 | 92 |
| Mié | 18 | 94 |
| Jue | 26 | 89 |
| Vie | 17 | 95 |
| Sáb | 14 | 96 |
FIG. 02 · TRACE TREE
Audit log
CID-9128low
CX WhatsApp
tool: CRM.lookup · latency 2.4s
CID-9132medium
Scoring
tool: Policy.check · latency 3.1s
CID-9135low
RAG Técnico
tool: Vector.search · latency 1.8s
CID-9141blocked
Outbound
tool: Consent.verify · latency 0.9s
FIG. 03 · TOKEN COST
Cost by model
FIG. 04 · CONTROL MAP
Framework mapping
| Control | NIST | ISO 42001 | OWASP | CSA | Status |
|---|---|---|---|---|---|
| AI Registry | GOVERN/MAP | §6 Planning | LLM05 | AIG-04 | Operational demo |
| Prompt Injection Filter | MEASURE | §8 Operation | LLM01 | AIS-02 | Active in chatbot |
| Human Review Queue | MANAGE | §9 Evaluation | LLM07 | AIG-09 | Realistic mock |
| Model Rollback | MANAGE | §10 Improvement | LLM10 | AIM-11 | Product roadmap |
| DPA + Subprocessors | GOVERN | §7 Support | LLM02 | DSP-01 | Template included |
CASES
Outcome-first demos for boardroom conversations.
Financial services
Contact center with human review
From 0 to 68% assisted resolution in six weeks, with a HITL queue for exceptions.
- 68% assisted resolution
- p95 < 3.2s
- 100% traced
Industrial B2B
Agentic sales pipeline
Lead prioritization, follow-ups and executive summaries synced to CRM.
- 23% more qualified pipeline
- 11h/week saved
- CRM hygiene +41%
Energy
Governed technical RAG
Document assistant with cited sources, role scopes and knowledge drift monitoring.
- 92% grounded target
- RBAC scopes
- Drift watch
PACKAGING
Clear entry points. Enterprise depth.
Executive Radiography
USD 950
5–10 business days · one real workflow
- One selected business workflow
- Up to 3 stakeholders
- Process map + baseline
- Systems, data, permissions and integrations
- Private architecture + estimated costs
- Risks, human approvals and rollback
- Configure/buy vs build alternatives
- Proceed, reformulate or stop + implementation budget
- Full USD 950 credit to first same-workflow implementation within 30 days
Governed Pilot
From USD 6,500
4–6 weeks + 30 days of operation
- 2–4 agents
- RAG
- CRM/WhatsApp
- Monthly evals
- Trust Core
Agentic Office
From USD 3,900/mo
monthly + usage + SLA
- SSO/SAML
- RBAC
- Separate environments
- Full Trust Core
- Optional 24/7 support
Enterprise Multi-model
Custom
license + usage + contract
- API estándar
- BYOK
- Tenant isolation
- Client branding
- DPA + subprocessors