How Circuito AI Works With Metso, ABB, Honeywell, FLSmidth, and Schneider — Not Against Them

TL;DR

Circuito AI is not a competitor to Metso, ABB, Honeywell, FLSmidth, or Schneider Electric. It is 100% DCS and OEM agnostic — an AI intelligence layer that integrates with whatever control system and equipment you already operate. Think of it as adding autopilot to a car that already has cruise control. Your DCS keeps the process stable. Circuito AI makes it intelligent.

The Question Every Mining CTO Asks

When we present Circuito AI to mining operations in Latin America, the first question is almost always the same:

“We already run Metso OCT” — or ABB 800xA, or Honeywell Experion, or FLSmidth PXP — “so what do we need this for?”

It is a fair question. These are serious platforms from serious companies. They cost millions to install, took months to commission, and your operators know how to use them. Nobody wants to hear that their existing infrastructure needs to be replaced.

Here is the answer: Circuito AI does not replace anything. It makes everything you already have work harder.

Cruise Control vs Autopilot

The analogy we use most often is cruise control versus autopilot.

Cruise control holds your car at a set speed. If the road goes uphill, the car slows down until the engine compensates. If traffic changes, you intervene manually. It is reactive, single-variable, and rule-based.

Autopilot reads the road ahead, adjusts speed before the hill, changes lanes when needed, and continuously adapts to conditions your cruise control cannot see.

Your DCS and its built-in APC are cruise control. They hold setpoints, follow rules, and respond to deviations. They do this well. But they cannot learn from yesterday’s ore to predict what tomorrow’s blend will need. They cannot optimize flotation recovery and grinding throughput simultaneously across hundreds of interacting variables. They cannot retrain themselves when a new mining zone sends different mineralogy through the plant.

Circuito AI is the autopilot layer. It reads thousands of process variables simultaneously, identifies non-linear relationships that no human operator or rule-based system can see, and writes optimized setpoints to your existing DCS every 10 seconds. Your DCS executes. Circuito AI decides.

What This Looks Like With Each OEM

Metso (OCT Platform)

What Metso provides: OCT with rule-based APC, fuzzy logic, and MPC. Proprietary sensors including RockSense, MillSense, and FrothSense for grinding and flotation monitoring.

The gap: Traditional APC that does not learn or adapt to changing ore conditions. Optimization capabilities are strongest with Metso’s own equipment, and the rule-based logic requires manual retuning when process conditions drift.

What Circuito AI adds: Reinforcement learning that works WITH Metso’s sensors and equipment. Circuito AI reads data from RockSense, FrothSense, and every other sensor on the OPC network. It uses that data — plus hundreds of other variables Metso’s rules do not consider — to write optimized setpoints back through the DCS. Metso’s sensors become inputs to a more intelligent optimization layer.

Nothing is removed. Metso’s sensors stay. OCT stays. The DCS stays. Circuito AI adds the AI that Metso’s rule-based APC cannot provide.

ABB (800xA / Expert Optimizer)

What ABB provides: Expert Optimizer with MPC, neural networks, and fuzzy logic on the 800xA DCS. Over 25 years of advanced process control experience across multiple industries.

The gap: Traditional MPC architecture — not modern deep learning. ABB’s cross-industry approach means their models are generic rather than mining-specialist. Performance is best on ABB’s own 800xA DCS.

What Circuito AI adds: Circuito AI integrates with 800xA AND any other DCS in the plant. It adds modern AI/ML that goes beyond legacy MPC — mining-specialist algorithms built by process engineers with 20,000+ production hours of AI-driven control in mineral processing. Where Expert Optimizer uses linearized models that degrade as ore changes, Circuito AI’s reinforcement learning continuously adapts.

If you run ABB 800xA at your concentrator, you keep running it. Circuito AI connects through OPC, reads your process data, and writes setpoints that ABB’s DCS executes. Your operators still use the same HMI. The difference is that the setpoints are now coming from an AI that understands your specific ore body and circuit dynamics.

Honeywell (Experion PKS / Forge APC)

What Honeywell provides: Forge APC with Profit Controller (MPC), Profit SensorPro (virtual sensors), and Plant-Wide Optimizer. Experion PKS DCS covering the full mineral processing chain.

The gap: Honeywell’s heritage is refining and chemicals — mining is a secondary focus. Their MPC is traditional, not AI-native. Published mining-specific optimization results are limited compared to their refinery track record.

What Circuito AI adds: Mining-specialist AI built by PhDs who work exclusively on mineral processing — not a refining APC adapted for mining. Circuito AI integrates with Experion PKS and reads data from Profit SensorPro virtual sensors. The key difference: Honeywell’s MPC assumes process relationships are approximately linear within operating ranges. Circuito AI’s neural networks capture the full non-linear behavior of flotation kinetics, grinding dynamics, and ore variability.

Your Honeywell investment is protected. Experion PKS continues to manage your process. Forge APC can continue running in parallel. Circuito AI adds a layer of intelligence that Honeywell’s mining APC, designed primarily for refinery operations, was not built to deliver.

FLSmidth (ECS/ProcessExpert)

What FLSmidth provides: ECS/ProcessExpert (PXP) for grinding and flotation APC. SmartCyclone and LoadIQ smart sensors. AI/ML integration beginning with PXP v8.5+.

The gap: AI/ML integration is still maturing. Optimization capabilities are tied to the FLSmidth hardware ecosystem. Implementation is complex, and results depend heavily on having FLSmidth equipment throughout the circuit.

What Circuito AI adds: Closed-loop AI optimization alongside FLSmidth equipment and sensors. No hardware dependency — Circuito AI delivers the AI layer that PXP alone cannot yet provide. SmartCyclone and LoadIQ data feeds directly into Circuito AI’s models, making FLSmidth’s smart sensors more valuable by connecting them to a broader optimization strategy.

FLSmidth is building toward AI. Circuito AI is already there. Until FLSmidth’s native AI matures, Circuito AI fills the gap — and when it does mature, Circuito AI’s vendor-neutral architecture means both can coexist.

Schneider Electric (EcoStruxure / Modicon)

What Schneider provides: EcoStruxure Process Expert IIoT platform with APC. Plant Advisor for AI self-learning. Modicon M580 ePAC native integration. Broad industrial coverage.

The gap: Mining represents roughly 15% of Schneider’s platform focus — they are not a mining specialist. Their optimization platform is generic, designed for water treatment, oil and gas, chemicals, and manufacturing alongside mining.

What Circuito AI adds: Purpose-built mineral processing optimization by mining PhDs. Circuito AI integrates with EcoStruxure infrastructure while adding the domain-specific AI that generic industrial platforms lack. Schneider’s platform is designed to manage and monitor. Circuito AI is designed to optimize the specific metallurgical outcomes that determine your concentrator’s economics.

Why Vendor-Neutral Matters

Most concentrators do not run a single vendor’s equipment. A typical LATAM operation might have a Metso SAG mill, ABB drives, an FLSmidth flotation circuit, and a Honeywell DCS. Each OEM’s optimization software works best with its own hardware. None of them optimize across the entire circuit because none of them see the entire circuit.

Circuito AI is different because it does not sell equipment. It has no hardware to favor, no sensors to push, and no DCS to protect. It connects to everything through standard industrial protocols (OPC-UA, OPC-DA) and optimizes the process as a unified system — from crushing to recovery.

This is the fundamental advantage of being DCS and OEM agnostic. Your process does not care which vendor made the mill, the flotation cells, or the DCS. The ore does not know. The physics does not know. Your optimization layer should not care either.

The Integration Architecture

For engineers evaluating the integration risk, here is how Circuito AI connects:

1. Data ingestion: Circuito AI reads process data from your existing OPC server — the same data your DCS historians already collect. No new sensors required (though Circuito AI can incorporate data from any additional sensors on the network).

2. AI processing: Circuito AI’s reinforcement learning models run on a dedicated on-premises server at your site. No cloud dependency. No data leaves your network. The AI processes thousands of variables simultaneously and generates optimized setpoints.

3. Setpoint delivery: Optimized setpoints are written back to the DCS through the same OPC connection. Your DCS executes these setpoints exactly as it would execute setpoints from an operator or from its own APC.

4. Operator override: Operators can override any Circuito AI setpoint at any time from the existing HMI. Circuito AI detects the override and adjusts. The operator always has final authority.

5. A/B verification: Every Circuito AI deployment includes A/B testing — periods of AI-on versus AI-off operation — to prove the measurable impact under your specific conditions.

The result: Circuito AI is fully operational within weeks, not months. There is no DCS migration, no hardware swap, and no retraining of operators on new interfaces.

What This Means for Your Existing APC Investment

If you spent two years and significant budget commissioning Metso OCT, ABB Expert Optimizer, or Honeywell Forge APC at your site, that investment is not wasted. Circuito AI does not ask you to decommission anything.

In fact, your existing APC makes Circuito AI’s job easier. A well-tuned APC provides stable base control — the cruise control foundation that Circuito AI’s autopilot builds on. The base layer handles the regulatory control loops and basic setpoint holding. Circuito AI handles the higher-order optimization that your APC was never designed to perform: multi-variable, non-linear, continuously adaptive optimization across the entire concentrator.

The question is not “Circuito AI or my existing APC.” The question is: “Am I extracting every possible dollar from my concentrator, or is my process control leaving money on the table?”

At every mine where Circuito AI has been deployed, the answer has been the same. Existing control systems were holding setpoints reliably. But they were not optimizing. Circuito AI found additional recovery, throughput, and energy savings that the existing APC could not — because rule-based and MPC systems were not designed to find them.

The numbers from production deployments: +2.9% throughput and +0.9% Ni extraction at Norilsk Nickel. +0.78% Cu recovery and -8% mill energy at Russian Copper Company. These gains came on top of whatever control systems were already running.

The Bottom Line

Circuito AI does not compete with Metso, ABB, Honeywell, FLSmidth, or Schneider Electric. It makes their platforms perform better.

Your DCS runs the process. Your sensors measure it. Your APC stabilizes it. Circuito AI optimizes it — continuously, adaptively, and across your entire concentrator as one integrated system.

If you are running any of these platforms at your LATAM mining operation and want to see what an AI intelligence layer can add, contact Circuito AI to discuss a pilot deployment. We will show you the gap between what your current system delivers and what your process is capable of.