How MICA and KorrAI Are Accelerating AI‑Enabled Mine Operations

The mining sector is standing on the brink of its most profound technological shift in decades. Since 2021, more than US $8.5 billion of investment and M&A activity has flowed into mining software, sensors, and automation solutions, with analysts forecasting the "smart‑mining" market to reach US $40–50 billion within the next ten years.

It was in June, at a MICA-organized event focused on AI Adoption in Mining Operations, that I saw the impetus in Canada’s mining sector to seize this opportunity. There is strong alignment between the government, private sector, and the people to continue investing in one of our key industries. After all, mining employs over 710,000 Canadians - direct and indirect - and is responsible for 5% of our GDP.

On Steadier Ground Thanks To MICA’s Catalytic Role

In 2023, MICA funded CA $0.99 million for KorrAI's “Satellite‑Based Mine Site Monitoring Using AI & InSAR” project. The funding helped accelerate research ideas into production‑grade tools and was critical for the development of several capabilities.

1. Accelerated InSAR Engine

• Re‑wrote and parallelized key portions of the open‑source STAMPS stack, cutting end‑to‑end processing from 7 days to ≈ 4.5 hours and scaling to >1 PB of imagery per year.
• Automated orbit/frame selection so new Sentinel‑1 scenes flow through the pipeline with zero human touch.

2. Integrated Sensor Network

• Deployed 9 GNSS stations, 5 extensometers, 14 radar corner reflectors across three flagship sites, streaming data to the cloud via rugged IoT loggers.
• Purpose‑built a data‑fusion service that time‑syncs sensor feeds with satellite passes, delivering millimeter‑accurate displacement curves.

3. AI Calibration & Early‑Warning Analytics

• Introduced an AI‑driven calibration layer that leverages in‑situ sensors to remove atmospheric noise from InSAR, raising accuracy to sub‑millimeter levels.
• Trained the “Severity Sense” LSTM / GRU model on five years of historical signals to forecast accelerating movement, shifting alerts from reactive to predictive.

4. KorrAI Explore Platform

• Launched a cloud‑native web app capable of streaming tens of thousands of deformation points through a custom vector‑tiling engine; users can pan, zoom, and query in real‑time even on low‑bandwidth connections.
• Added 3‑D terrain views, multi‑sensor overlays, and SOC‑2 compliant security so field crews and executives share a single source of truth.

5. Field‑Proven at Scale

• Rolled out pilots at five Canadian mines including Eldorado Gold, Agnico Eagle, EDM Resources, and TransAlta, operating year round including in harsh Canadian winter conditions to validate robustness.
• Early warnings proved their worth: one rainfall event triggered an alert for anomalous movement that led to an inspection and mitigation before erosional damage could escalate.

Bottom line: The first MICA funding moved KorrAI from lab prototype to TRL‑9, production‑ready platform in under 24 months, setting the stage for the next leap.

“MICA is proud to have supported Korr AI and is excited by the tangible results they have been able to achieve as a result of their project. The solution they’ve developed solves a significant challenge in the mining industry which is the monitoring of critical pieces of infrastructure, allowing for proactive maintenance.”, says Charles Chamirai Nyabeze who is Network Director of MICA.

Why the Next Leap Must Be AI‑Native

While high‑resolution monitoring has become indispensable, the industry’s bigger hurdle is knowledge work: synthesizing terabytes of geospatial data, inspection logs, and regulatory documents into clear, timely decisions.

Throughout several conversations with engineering teams and executives, we’ve identified recurring challenges:

• Fragmented data silos across sensors, reports, and legacy databases
• Rising ESG and tailings‑governance scrutiny (e.g., GISTM compliance)
• Talent shortages as seasoned geotechs retire faster than new specialists can be trained.

Meeting these challenges demands an AI layer that is mining‑aware by design, not a generic chatbot.

“Data in the mining industry is readily available. However, making it accessible and actionable continues to be a hurdle many operators are faced with. Korr AI and their TRAIL platform are looking to bridge that gap, and MICA is looking forward to supporting the development of the solution.”, says Kyle McCall, the Director of Commercialization Services at MICA.‍

Introducing TRAIL: An AI Co‑worker for Mining Operations

Backed by a new round of MICA funding in 2025, KorrAI is building TRAIL, an industry-grade AI co-worker that is fluent in geotechnical ontology and understands the specifics of mine assets it’s deployed to.

What sets TRAIL apart?

The first and second wave of Generative AI products, including the ones from foundational model creators, lack the ability to give the AI rich context. More importantly, the AI models aren’t primed for working on mining operations.

TRAIL is purpose-built for the industry and is equipped to deliver insights that are accurate, grounded, and verifiable.

Converting Data Streams to Decision Support

We want every field engineer, geotech, and executive to have near real-time situational awareness.

By layering TRAIL on top of the sensor streams, cloud storage for document and reports, email inboxes, inspections logs, and other data sources, we’re making insights a query away. Moreover, TRAIL is engineered to be proactive.

From being able to perform complex analysis, to flagging accelerating movement or recommending prioritizing inspections, and even drafting regulatory filings, TRAIL is an AI co-worker that plugs into the daily workflow.

Looking Ahead: Canada’s New Export Opportunity

My co-founder, Rob McEwan, had the opportunity to present TRAIL at the aforementioned event. In his introduction, he alluded to the opportunity ahead of us - Canada as an exporter of AI co-workers.

As mining operations deepen and ESG expectations intensify, AI co‑workers will move from novelty to necessity. I’m of the opinion that Canada should lead the way, so as to propagate the Canadian know-how, best practices, and excellence we’re known for. With MICA’s continued partnership, KorrAI is committed to ensuring that every mine - regardless of size or geography - can harness the full power of earth observation, knowledge graphs, and large‑language models.

Btw, here's Rob presenting at the conference.