Predict. Act. Prove. · Predictive intelligence for heavy industry

Your operation can warn you before it fails.

AMARU is a predictive intelligence company for heavy industry. One agentic AI layer unifies fixed sensors, third-party systems and robots with the platforms your plant already runs — to anticipate the failure, execute the protocol automatically and turn every action into evidence the regulator accepts — while robots take on the hazardous exposure and people take command.

24/7
Robots and AI on continuous rounds and vigilance — people out of the risk area, in command of decisions
< 5 min
for a critical event to be detected, prioritized and escalated — with the response protocol already in motion
100%
of actions recorded on an immutable audit trail — regulator-ready evidence, mission by mission
1
single intelligence layer over robots, sensors and the systems your plant already runs
The Great Shift

The market changed. Watching is no longer enough.

For decades, running heavy industry meant reacting well: an operator in front of SCADA screens, three shifts a day, activating protocols by hand after an indicator left its range. Three changes just made that model obsolete.

01

The signals outgrew human attention

SCADA, fixed sensors, cameras, maintenance systems: a modern plant produces more critical signals than any team can watch. Reactive vigilance does not fail for lack of effort — it fails by design. A deviation missed for minutes can compromise an entire production run.

02

AI learned to predict — and to act

Agentic AI no longer just triggers alarms after the limit is crossed. It reads trends, anticipates the deviation and executes the response protocol on its own — escalating to humans only the decisions that truly demand judgment.

03

The regulator now demands continuous proof

Safety regulation across the region — Peru's mining code first among it — is precise: gases measured, airflow verified, fans monitored, records kept. Point-in-time readings and paperwork are becoming a liability; continuous, auditable evidence is becoming the license to operate.

Separately, these changes look small. Together, they create a new way to operate.
The Answer

AMARU was born from this convergence.

When signals overflow, when AI can act, and when the regulator demands continuous proof, the answer is not another robot or another screen. It is an intelligence layer above all of them. That is what we build.

AMARU is not a robot company. It is not a distributor. It is an intelligence company — the platform that takes your operation from reactive to predictive: Predict. Act. Prove.

The Human Factor

Robots go into the risk so people don't have to.

Heavy industry still pays its hardest price in people: toxic gases, extreme heat, explosive atmospheres and unstable ground put lives at risk every day, on routine rounds. Our intelligence exists to change who takes that exposure: robots assume the dangerous work, 24/7 and without fatigue, while people move to the NOC, to analysis and to command. Fewer people exposed means fewer accidents possible.

The dangerous round stops being human

Inspections in gas-laden galleries, around furnaces and inside classified areas are taken over by robots — continuous presence in exactly the places where no person should linger.

A routine reading should never be a risk

Every worker taken out of a hazardous area is one less exposure in your operation — and one more family at ease at home. That is the metric behind every solution we build.

From exposure to command

Operators and guards don't disappear — they move up: to the NOC, to supervision, to decisions. The AMARU Academy requalifies today's teams to run tomorrow's predictive operation.

Predictive intelligence with a human purpose: the predicted failure protects production; the robot in the field protects lives.
Predictive Solutions

We don't sell monitoring. We sell what happens before the failure.

Four solutions, each born from a real operational pain and a verifiable regulatory or production requirement — and each presented by what it predicts, not just by what it measures. Delivered over your existing sensors, our robots, or both.

Underground Ventilation & Gases

Production that doesn't stop

Continuous monitoring of gases (CO, CO₂, NO₂, O₂), air velocity and airflow, plus telemetry from the main fans — predicting fan failure and air-quality degradation before the deviation, not after it.

Result: continuous, auditable compliance, reports ready for inspection and maintenance scheduled before the stoppage.

Thermal Intelligence for Smelters & Furnaces

The end of reactive vigilance

24/7 thermal signature of furnaces and critical equipment, with trend analysis that catches the deviation before it leaves the operating range — and agentic AI that activates the protocol automatically.

Result: production protected, protocols executed in seconds, and no one condemned to watch screens across three shifts.

Inspection as a Service

Wear detected before the break

Recurring rounds with LiDAR, thermal imaging and computer vision — mobile sensing that reaches where fixed sensors can't, predicting support deformation, corrosion and wear before collapse or breakdown.

Result: comparative 3D mapping, an auditable history of every asset and maintenance prioritized by real condition.

Security as a Service

Risk detected early

Autonomous 24/7 patrol of perimeter and critical areas, with anomaly and risk-pattern detection across cameras, access systems and mobile robots — orchestrated by the NOC.

Result: orchestrated response in minutes and evidence with a complete audit trail — for your security team and your insurer.
Focus discipline: underground ventilation and smelters are our first showcases — both backed by explicit regulatory demand and real operating cases. The predictive capability of each solution is validated progressively on your asset before it scales. No inflated promises.
The Regulatory Anchor

Born to meet each country's regulation — proven first against Peru's mining code (D.S. 024-2016-EM, arts. 246–251)

Art. 246Gas measurement (CO, CO₂, NO₂, O₂) at the start of every shift, with records kept
Continuous monitoring with auditable records — beyond the point-in-time reading
Arts. 247–248Minimum airflow per person and air velocity within regulated ranges, by altitude
Continuous verification of airflow and velocity, calibrated to the operation's altitude, with non-conformity alerts
Art. 249Main fans with automatic alarms, redundant power and maintenance to the maker's plan
Fan telemetry and predictive management of the maintenance plan
Art. 250A fan failure means stopping the workings and closing access until air quality is verified
Failure predicted, maintenance before the stoppage — production that doesn't stop
Art. 251Large fans require control panels monitoring operation, gases and temperature, with alarms and backup power
Platform & NOC as the 24/7 monitoring layer, with agentic AI and an immutable audit trail
Art. 250 is the economic argument: a stopped fan is a stopped mine. We don't sell inspection robots — we sell continuous, auditable compliance and production that doesn't stop. And Peru is the starting point, not the boundary: the same engine maps each country's requirements — Brazil's NR-22, Chile's DS 132 and beyond — into continuous, auditable evidence, over the sensors you already have.
Gases, airflow and fan telemetry unified into one predictive picture of the underground.
Application · Underground Mining

The ventilation that warns before it stops the mine

In underground operations, air is the license to produce. Yet most mines still measure gases once per shift and discover fan trouble when the alarm goes off — exactly when regulation forces the workings to stop. We connect fixed gas sensors, fan telemetry and mobile robot readings into one predictive layer.

  • Continuous reading of CO, CO₂, NO₂ and O₂ with auditable records — every gallery, every minute
  • Airflow and air velocity verified against the regulated thresholds for your altitude
  • Predictive maintenance of the main fans — the failure anticipated, the stoppage avoided
  • Robots as mobile sensors where fixed instrumentation cannot reach
  • Inspection-ready reports generated automatically, mission by mission
Thermal deviation anticipated by trend — the protocol in motion before the range is crossed.
Application · Smelters & Furnaces

The furnace watched by trend, not by shift

Smelting furnaces run at extreme temperatures under human vigilance: an operator in front of SCADA around the clock, alerted only when an indicator leaves its range — and every protocol activated by hand. In that model, one undetected rise can compromise an entire production run. We replace reactive watching with a 24/7 thermal signature and trend analysis that anticipates the deviation.

  • Continuous thermal signature of furnaces and critical equipment, across fixed and mobile sensing
  • Trend analysis that flags the deviation before it leaves the operating range
  • Response protocols executed automatically by agentic AI — critical decisions escalated to your team
  • Integration with the SCADA and historian you already run — no rip-and-replace
  • Every event, action and outcome recorded on an immutable audit trail

Energy & Utilities — roadmap

Thermal rounds of substations and transmission assets, predictive inspection of thermal and renewable plants — the same intelligence layer, extended as the showcase cases are validated.

Ports, Oil & Gas & Heavy Industry — roadmap

Terminal yards, tank farms, classified areas and process plants — with Ex-certified platforms and the same rule: prediction is only promised where it has been proven.

The Commercial Journey

From first conversation to continuous operation

STAGE 1

Awareness

An executive session on your operation's terms: where reactive vigilance is costing you, what the regulation demands, and what predictive operation would look like on your assets. No commitment, no generic demo.

STAGE 2

Diagnosis

We map critical assets, existing sensors and systems (SCADA, historian, CMMS) and your current baseline — and return a business case with measurable success criteria before any contract.

STAGE 3

Implementation

A paid pilot on one critical asset: platform integrated, predictive capability validated against your own baseline, results measured. Over your sensors, our robots, or both — with credit convertible into the contract.

STAGE 4

Continuous Operation

The NOC takes over 24/7: prediction, automatic protocols, regulator-ready reporting, SLAs and a dedicated relationship executive — expanding asset by asset as value is proven.

No rip-and-replace. No mandatory CAPEX. No unproven promises. Each stage exists to remove one risk from your decision.
Platform & NOC

The core of everything we deliver

The platform is the product. Robots and sensors — ours or yours — are its eyes and hands. One agentic AI layer ingests every source your operation already has, normalizes it into a single operational picture, predicts deviations by trend, executes routine protocols autonomously and escalates safety-critical decisions to humans. Operated from our Network Operations Center, 24/7.

Inputs

  • Fixed & third-party sensorsgas, thermal, vibration, LiDAR — including the ones already installed
  • Robotsmobile sensor and executor — ours or the client's
  • Your plant systemsSCADA, historian, CMMS, ERP, CCTV, access control, fire & gas

Agentic AI NOC

  • Normalizesheterogeneous sources reduced to one operational picture
  • Predictstrends and signatures read before the deviation leaves the range
  • Actsprotocols executed autonomously — safety-critical calls escalated to humans
  • Provesevery action on an immutable, regulator-ready audit trail

Outcomes

  • Productionfailures anticipated, maintenance before the stoppage, less downtime
  • Compliancecritical alerts in under five minutes; reports ready for inspection
  • Peopleout of hazardous exposure, in command of the decisions that matter
Before & After

From reactive to predictive — a different operating model

A human watching screens, three shifts a day
Agentic AI watching every signal — humans deciding only what is critical
Alarm after the limit is crossed
Deviation predicted by trend, before it leaves the range
Protocols activated by hand
Protocols executed automatically, with defined escalation
Gas readings once per shift
Continuous monitoring with auditable records
Compliance as paperwork after the fact
Regulator-ready evidence generated while the operation runs
Predict. Act. Prove. — not a better version of monitoring. A different way to operate.

The integrated ecosystem — hardware is the entry point; intelligence is the product

Unitree
B2
The volume workhorse

6 m/s top speed · >40 kg payload · 5 h endurance · IP67 · open SDK — continuous rounds at fleet scale.

Deep Robotics
X30
The severe-environment specialist

−20 °C to +55 °C · 45° slopes & industrial stairs · IP67 · sees in total darkness — born for mines and tunnels.

ANYbotics
ANYmal X
The hazardous-area authority

Ex-certified for explosive atmospheres — inspection where no uncertified device may enter.

We integrate the right platform for each task and environment — and, above all, the fixed sensors and systems your plant already owns.

Certified payloads & sensing

MSA Safety gas detection Axis thermal & optical vision Ouster LiDAR & 3D mapping
Commercial Models

Three layers. One principle: intelligence as recurring revenue.

Engagement models designed for how large operations actually buy — anchored on a recurring intelligence subscription, made flexible by asset and usage levers, and built to fit the systems you already run.

Layer 1 · The Recurring Core

The revenue of today — and the relationship of every month:

  • Platform & telemetry as a subscription (SaaS) — including for robots and sensors you already own
  • Robot-as-a-Service with SLA tiers — no CAPEX required from the client
  • Complete solutions: hardware, payloads, integration engineering and commissioning
  • Paid pilots with success criteria and credit convertible into the contract

Layer 2 · Flexibility & Asset Levers

Pricing that adapts to your operation — on an asset base that strengthens ours:

  • Pay-per-use (robot-hour) for operations with variable demand
  • Take-or-pay floors and leasing to unlock the client's CAPEX
  • Robots and sensors can be AMARU-owned assets — a fleet that grows the company's asset base and value, and lets you consume the service with flexibility, independent of the intelligence subscription

Layer 3 · Ecosystem & Partnerships

Integration engineering that builds the moat:

  • Certified integrations of robots, payloads and sensors into a single platform
  • Local service partners connected to our business intelligence — field presence close to your asset
  • Evolutionary models — shared fleets, white-label, data products — enter the roadmap only after market validation
One rule inherited from our business plan: nothing enters a financial projection — or a promise to you — without validation.
Why AMARU

Six differentiators, each tied to a result of yours

Methodology

Predict. Act. Prove. — with predictive capability validated progressively on your asset, against your own baseline. For you: promises you can take to the board.

Technology

Agentic AI over total integration — SCADA, historian, CMMS, ERP, CCTV, fire & gas, fixed sensors and robots. For you: one picture that doesn't just inform. It acts.

Team

Operators of heavy industry, risk & compliance, data and technology platforms across Latin America. For you: a partner that speaks the language of your operation — and of your regulator.

Processes

Human-on-the-loop for safety-critical decisions, one single safety standard and an immutable audit trail on every action. For you: governance your auditors accept.

Service & Relationship

A 24/7 NOC, local service partners integrated into our intelligence, and a dedicated relationship executive per account with periodic value reviews. For you: response close to the asset — and one person who answers for everything.

Innovation

A platform that evolves monthly: new predictive models, new certified payload kits, new integrations. For you: the solution gets better over time — without a new project each time.

Who We Are

Named after the Andean guardian. Built to anticipate.

Amaru is the guardian serpent of the Andes — a symbol of protection and foresight. We were born watching heavy industry operate the other way around: control rooms where people watch screens around the clock, alarms that only speak after the limit is crossed, protocols that depend on someone's reaction, and compliance assembled as paperwork. We decided to solve exactly that problem: the reactive operation.

Our vision is a Latin American heavy industry that operates predictively — where failures are anticipated, protocols execute themselves and every action becomes evidence. Our purpose is to protect what matters in an operation: people, production and the license to operate. Our mission is to turn the signals a plant already produces into anticipation, action and proof.

That is why AMARU is not a robot company, nor a distributor — it is an intelligence company. Hardware is the entry point; the platform and the NOC are what we deliver. We operate as one continental platform with deep local roots, starting in Peru and Brazil, with a staged roadmap across the rest of Latin America.

Predictive by principle

Every solution is presented by what it predicts — and every prediction is validated on the client's asset before it scales.

Evidence over promises

Paid pilots measured against your baseline, no number without a source, no capability claimed before it is proven.

No hardware lock-in

The platform comes first: we never lock a client into hardware, because hardware is not the product.

People in command

AI executes the routine; humans decide the critical. Safety-critical calls always escalate to people — yours and ours.

“We believe no operation should learn about a failure after it happens — and no one should risk their life doing work a robot can do. The future of heavy industry will not belong to whoever has the most screens — it will belong to whoever predicts. We are building that future in Latin America, and we invite you to operate in it first.”
The AMARU LATAM manifesto

Built by operators who know the terrain

Samuel Reis

Samuel Reis

Chief Executive Officer

20+ years in IT, consulting & data intelligence — Andersen, Booz Allen, EMC, Informatica. MSc in AI. Ran P&L above US$ 50M.

Eduardo Barboza

Eduardo Barboza

Chief Operating Officer

Managing partner of LATiNFORMATICA, São Paulo. 25+ years building technology businesses in Brazil. Leads LATAM operations & the Manaus depot.

Felipe Maia

Felipe Maia

Chief Financial Officer

12+ years leading FP&A, corporate finance & operations in technology, SaaS and fintech. MSc · MBA. Treasury and multi-country structuring.

Carlos Chagas

Carlos Chagas

Chief Product Officer

Digital-transformation leader across Brazil, Mexico & the US. Founder, Growth IT Consulting; co-founder, ViViTECH. Owns the NOC product & payload kits.

Wendy Rivera

Wendy Rivera

Chief Strategy Officer

20+ years in risk, compliance & anti-fraud for energy & mining. Industrial Engineer (U. Lima), MBA (PAD). ISO 31000 / 37001 · ACFE. Leads strategy & the Peru front.

Where we operate

🇵🇪 PeruLaunch market · Lima
🇧🇷 BrazilLaunch market · São Paulo & Manaus
🇲🇽 MexicoExpansion roadmap
🇨🇱 ChileExpansion roadmap
🇨🇴 ColombiaExpansion roadmap
Central America & CaribbeanExpansion roadmap
ESG · Safety · Quality

Impact built into the operating model

The return that never shows up on a spreadsheet is the failure that never happened — and the person who never had to be there. Our impact model is structural, not bolted on.

SOCIAL

People out of harm's way

Predictive operation takes routine exposure out of hazardous areas — and the AMARU Academy requalifies today's operators and guards to run tomorrow's predictive operations.

ENVIRONMENTAL

Incidents prevented, not remediated

Electric fleets and a platform whose core job is anticipating environmental incidents — gases, thermal deviations, structural wear — before a failure becomes a disaster.

GOVERNANCE

Governed for trust

An immutable audit trail on every action, human-on-the-loop for critical decisions, safety committees with an external seat and data protection to the strictest standard in the region.

Relationship

Start with a diagnosis. Prove it on your asset. Operate in the predictive.

Tell us about your operation — the assets that cannot stop, the signals nobody manages to watch, the reports the regulator demands. We will come back with an operational diagnosis and a pilot proposal measured against your own baseline. And as a client you get what large operations require: a dedicated relationship executive, periodic value reviews and a direct line to the NOC.

Commercial & diagnosiscontact@amarulatam.com
OfficesLima, Peru · São Paulo, Brazil
Continuous relationshipA dedicated executive per account and quarterly value reviews
We reply within two business days.