What Split Engineering Does

Split Engineering, a part of Hexagon, is a specialized technology provider serving the global mining and metals industry. Founded in 1997 and based in Tucson, Arizona, the company has established itself as a leader in rock fragmentation analysis. Its core business revolves around providing software and hardware solutions that measure, analyze, and optimize the size distribution of rock fragments after blasting. This data is critical for mines, as fragmentation directly impacts the efficiency and cost of every downstream process, from loading and hauling to crushing and grinding. By turning visual and sensor data from blast sites into actionable insights, Split Engineering helps its clients maximize mineral recovery and reduce energy consumption.

Why AI Matters at This Scale

For a company of 500–1000 employees operating within a large, innovation-focused parent like Hexagon, AI represents a strategic imperative to deepen its value proposition and defend its niche. At this midsize scale, the company has accumulated decades of proprietary, domain-specific data but likely lacks the vast R&D budgets of tech giants. AI offers a force multiplier, enabling Split Engineering to automate complex analyses, uncover hidden predictive patterns, and develop next-generation products that move beyond descriptive analytics to prescriptive and predictive intelligence. This evolution is crucial for retaining and expanding its client base in a sector under relentless pressure to improve margins and operational sustainability.

Concrete AI Opportunities with ROI

1. Predictive Fragmentation Modeling: Developing machine learning models that predict fragmentation outcomes based on blast design parameters, rock geology, and historical performance. This allows mines to simulate and optimize blasts before they occur, potentially reducing oversize material (which requires expensive secondary breaking) and fines (which cause processing issues). The ROI comes from increased throughput in crushing circuits and significant savings in explosive energy and downstream processing costs. 2. Real-Time Ore Grade Sensing via Computer Vision: Implementing AI-powered image analysis on photos or video feeds of muck piles to provide instant, preliminary estimates of ore grade. This enables dynamic adjustment of haul truck routing and mill feed blending, maximizing the value of material sent to the processing plant. The impact is direct: higher recovery rates and reduced dilution, leading to increased revenue per ton mined. 3. Autonomous Drilling Analytics: Using AI to analyze real-time data from drill rigs (vibration, pressure, torque) to optimize drilling parameters automatically. This can extend drill bit life, improve penetration rates, and ensure blast hole consistency. For clients, the ROI manifests as lower consumable costs, reduced downtime for bit changes, and more consistent blast results, directly lowering operating expenses.

Deployment Risks for a 500–1000 Person Company

The primary risks are not purely technological but organizational. Integration Complexity: Embedding AI into legacy on-premise or ruggedized field software suites can be challenging, requiring careful API design and potentially slowing deployment. Skill Gap: While Hexagon may provide resources, the specific team may lack deep ML ops and data engineering talent, risking project delays or underperforming models. Client Trust & Explainability: Mining engineers are rightfully skeptical of "black box" recommendations. Failure to build transparent, explainable AI that aligns with their deep domain expertise will lead to rejection, regardless of algorithmic accuracy. Data Silos & Quality: Valuable data may be trapped in isolated client deployments or in inconsistent formats across decades of projects, requiring significant upfront effort to curate a unified training dataset. Navigating these risks requires a focused pilot strategy, strong partnerships with client-side champions, and a commitment to building AI that augments, rather than replaces, human expertise.