AI-powered onboarding and operational intelligence for mining.
Your mine is complex. Bringing people up to speed doesn't have to be. MineTwin connects to your operational systems and builds a living intelligence layer - one AI guide that knows every role, every asset, and every answer.
A live digital twin of your site, an expert agent that reads across every operational system, and a console that runs the whole onboarding pipeline. Three surfaces, one platform, every role.
Role-aware 3D induction with an AI guide that knows the site, the assets, and the script. Replaces the classroom, the PowerPoint, and the awkward windshield tour.
Ask hard operational questions across fleet telemetry, CMMS, environmental, and financial systems. Every answer cites its sources, declares its confidence, and respects role-based access.
Run the whole onboarding pipeline from one screen. Assign, watch, nudge, sign off. Drop into any live session in the twin to assist. Compliance numbers in the sidebar, never on a spreadsheet.
Every answer is grounded in your operational data, cites its sources, declares its confidence, and respects role-based access.
Select your role and we'll show you the content, data, and value proposition that matters most to you.
Induction is mandatory. Consistency hasn't been.
Consistent induction used to mean dedicated headcount. MineTwin is the headcount.
At scale, induction quality varies by site, by rotation, by whoever from HR, health & safety, or operations was available that day. MineTwin connects to your operational systems and understands your mine in context - more completely than any individual ever could, because no single person has full visibility across all of it at once. The same standard. Every person. Every site.
Induction is mandatory. Consistency hasn't been.
Consistent induction used to mean dedicated headcount. MineTwin is the headcount.
For the first time, the knowledge your experienced people carry - the site, the equipment, the history, the context - lives in a platform instead of people. Induction has always depended on who was available. It doesn't anymore.
$1.8M per site per year in direct labor waste. 800+ qualifying mines globally. A founding team that already built this at enterprise scale inside one of the world's largest mining companies. We're productizing a decade of enterprise mining technology at $100K per year.
2,350 induction events × 6 hours × loaded staff costs. That's one site. Most operators run dozens. It's a known number. It gets budgeted. It doesn't get fixed - because until now, there was no mechanism to fix it.
New workers on unfamiliar sites. Generic briefings that didn't stick. Role-specific hazards that weren't covered. The incident rate in the first three months is a direct consequence of the quality of the induction that preceded it.
They arrive qualified. They leave trained - eventually. The gap between arrival and full productivity is a year of reduced output, elevated supervision cost, and institutional friction. Multiply that across every new hire, every rotation, every site.
When your senior metallurgist retires, her 30 years of process knowledge retires with her. When your best pit supervisor moves on, every workaround he knew leaves too. The organisation keeps running. It just runs a little worse. Every time.
Inductions pull people from across the operation - HR, health & safety, and operations all contribute. The content is standardized in slide decks. But who delivers it, and how well, depends entirely on who's available that rotation. Every group loses hours they should be spending on their actual job.
The haul truck driver and the instrument tech sit through the same session. Different equipment, different hazards, different SOPs - same content. Not because anyone chose to do it that way. Because there was never a cost-effective way to do it differently.
You've seen both ends. The H&S lead who covers everything thoroughly because she genuinely cares. The operations person who rushes through it because there's a breakdown on the ramp and his radio hasn't stopped. Same slide deck. Completely different outcomes.
They know the formal content. They don't know your site's specific quirks - which equipment has the recurring fault, which road gets icy first, who to actually call when something goes wrong at 3am. That knowledge takes months to absorb. And it comes at the expense of the experienced people who have to answer the same questions every rotation.
2,350 events × 6 hours × 1.5 staff per event. That's the direct labor behind the headline number. Each induction pulls from HR, health & safety, and operations - three groups losing hours from their actual roles to deliver the same standardized slide deck, every rotation.
Remote or FIFO operations, 300–1,000 employees, 30%+ annual turnover. That's the addressable market for Tier 1 alone. Every one of them is running this problem right now. None of them have a software solution for it.
Trimble, Hexagon, and Bentley sell asset registers and maps. None of them know who the workers are. Generic LMS platforms don't know the mine. No one has connected operational data, 3D spatial context, and role-aware AI into a single induction and knowledge platform. The category is open.
The reason no one has built this is not that the problem isn't real. It's that the technology - reliable real-time AI, industrial 3D digital twins, affordable cloud infrastructure - only matured enough to make it possible in the last 18 months. The timing is not accidental. It's a window.
MineTwin connects to the systems you already run - your CMMS, ERP, HR platform, operational data - and builds a living intelligence layer on top. One AI guide. Every person. Every role. Every question.
The moment someone logs in, the system already knows their name, their role, their crew, their supervisor, their equipment assignment, and their certifications. The mine knows who they are before they ask a single question.
A haul truck operator gets the pit, the haul road network, and the ROM hopper. A metallurgist gets the flotation circuit, the reagent regime, and the lab. Same platform. Completely different experience. Delivered consistently, every rotation, without pulling anyone off the floor.
And it doesn't stop after orientation. The AI guide stays available - answering questions, surfacing data, explaining systems - for as long as the person is on site. The new shift supervisor at 2am who needs the fault history on a specific truck gets an answer in seconds instead of waking someone up.
The new mining engineer who used to need a year to understand what's actually going on has a tool that has been running on this operation for months already. It knows the mine. Now they can too.
Under the hood: MineTwin's data engine ingests raw site data, cleans and categorizes it, stages and indexes it for real-time retrieval, and structures it into the Digital Twin knowledge graph that powers everything. The intelligence layer doesn't sit on top of a database - it understands the relationships between equipment, people, processes, and history.
How the Expert Agent works: When your shift supervisor asks a question at 2am, the system checks pre-computed views first - 23 verified answers covering 80% of common questions, returned in under 10 seconds. If the views don't cover it, a SQL Agent converts the question to a live query against Databricks. It then generates testable hypotheses and checks them against the evidence. Finally, it synthesizes everything into a structured response: the answer, the data, an interpretation, and a recommended action. Every number is traceable - the agent never fabricates data.
Role-based access control: The system knows who's asking. A General Manager sees everything - financials, personnel, failure history. A Maintenance Superintendent sees their department's costs and equipment data. A Haul Truck Operator sees production and safety data relevant to their area - no financials, no personnel records. We have 9 role profiles covering 150+ job titles across the mining industry. This filtering happens at the AI level, not just the UI - if a role doesn't permit it, the data never reaches the user.
| Asset | 12-mo Cost | Util % | Status |
|---|---|---|---|
| TRK-777-06 | $847K | 71% | Critical |
| TRK-777-03 | $412K | 86% | Normal |
| TRK-777-11 | $389K | 83% | Normal |
| TRK-777-08 | $371K | 84% | Normal |
| Fleet avg | $368K | 84% |
Role, crew, shift, supervisor, certifications, access zones - loaded at login. The truck driver and the electrician are in completely different mines from the moment they arrive.
Not a diagram. A real-time digital twin of your mine - every asset placed, every relationship mapped. Ask a question and the 3D model highlights exactly what you're asking about.
Four orientation durations - 5, 10, 20, and 60 minutes. Narration scripted with mining professionals, Q&A is live. After orientation, the AI stays - equipment history, SOPs, production data, cost centers - accessible any time.
Invitations sent to personal email. Inductees complete orientation before they set foot on site - reducing time on the ground and getting people productive from day one.
MineTwin connects to the systems a mine already runs - CMMS, ERP, HR platform, operational data - and builds a living intelligence layer on top. One AI guide per site. Every worker. Every role. Every question.
The moment a worker logs in, the system already knows their name, role, crew, supervisor, equipment assignment, and certifications. The mine knows who they are before they ask a single question.
A haul truck operator sees the pit, the haul road network, and the ROM hopper. A metallurgist sees the flotation circuit, the reagent regime, and the lab. Same platform. Completely different experience. Delivered consistently, every rotation, without pulling anyone off the floor.
The AI guide doesn't stop after orientation. It stays available - answering questions, surfacing data, explaining systems - for as long as the worker is on site. A shift supervisor at 2am who needs the fault history on a specific truck gets an answer in seconds instead of waking someone up.
A new mining engineer who would normally need a year to get up to speed has a tool that has been running on that operation for months already. It knows the mine. Now they can too.
Under the hood: MineTwin's data engine ingests raw site data, cleans and categorizes it, stages and indexes it for real-time retrieval, and structures it into the Digital Twin knowledge graph that powers everything. The intelligence layer doesn't sit on top of a database - it understands the relationships between equipment, people, processes, and history.
How the Expert Agent works: The AI checks 23 pre-computed views first - verified answers covering 80% of common questions, returned in under 10 seconds. If the views don't cover it, a SQL Agent converts the question to a live query. It then generates hypotheses, tests them against evidence, and delivers a structured response with traceable data, confidence levels, and recommended actions. The agent never fabricates data - if it doesn't have the answer, it says so.
Enterprise-grade access control: 9 role profiles covering 150+ job titles across the mining industry - from GM to contractor to apprentice fitter. The system maps any job title to the right access level automatically. This is governance built into the AI, not bolted on. If a role doesn't permit access to financial or personnel data, the AI itself redacts it - the data never reaches the user.
| Issue | Impact | Period |
|---|---|---|
| Trades turnover | ~12% capacity loss | 18 months |
| Reagent cost creep | $340K overspend | 6 months |
| Haul cycle drift | $2.1M/yr equiv. | Last quarter |
Role, crew, shift, supervisor, certifications, access zones - loaded at login. The truck driver and the electrician see completely different mines from the moment they arrive.
Not a diagram. A real-time digital twin of the mine - every asset placed, every relationship mapped. Workers ask a question and the 3D model highlights exactly what they're asking about.
Orientation ends. The AI doesn't. Four orientation durations - 5, 10, 20, and 60 minutes - from quick site overview to full induction. Narration is scripted with mining professionals. Q&A is live - the AI answers using the same Expert Agent with role-appropriate filtering. Completion is logged automatically.
Designed with security protocols so invitations can be sent to personal email. Inductees complete a significant portion of orientation before they ever set foot on site - reducing time on the ground and accelerating productivity.
Behind the 3D interface is a complete operational data architecture - MineZero - our production-scale proving ground - demonstrates the data ingestion patterns across 4.2 million records and 68 Databricks tables. Fleet telemetry, maintenance work orders, safety incidents, workforce profiles, and financials - the same architecture deploys to your site. Ask it anything about your mine.
Your analytics team produces reports. Those reports cover what you already know is wrong. They land on your desk two weeks after the data was generated, formatted for a meeting that happened last Thursday.
MineTwin answers the questions you haven't thought to ask yet. It looks across every domain - maintenance, production, safety, workforce, costs - and surfaces the correlations that don't show up in any single report.
The VP of Operations who asks "where am I losing money I don't know about yet" gets three specific answers, with the data to back them, in under ten seconds. That's not a dashboard. That's an analyst who knows your mine as well as your best people do - and never goes home.
This is what every site you operate could have. The data is already there. MineTwin builds the intelligence layer on top of it.
How the data gets in: Your mine already has data systems - a fleet management system like MineStar or Wenco, a CMMS like SAP or Maximo, environmental monitoring stations, your HR system, your financial system. That data flows into a Databricks lakehouse - a single data warehouse that brings everything together. We organize it in three layers:
Reference data - your equipment register, employee roster, PM plans, environmental limits. The master data that defines the operation.
Operational data - daily production, fleet telemetry, work orders, incidents, environmental samples. What's happening right now.
Pre-computed views - 23 common questions already answered and verified. Safety trends, fleet costs, production actuals vs plan. 80% of the time, the answer is already computed and returned in under 10 seconds.
The data then surfaces in three ways: through the 3D environment when you click on an asset, through the Expert Agent when you ask a question in natural language, and through the onboarding agent during guided 3D tours for new employees.
Behind the 3D interface is a complete operational data architecture - production records, maintenance history, safety events, workforce data, and financials. The system answers real operational questions about any mine it's deployed on.
Mine operators today rely on analytics teams that produce static reports - covering what's already known, two weeks after the data was generated, formatted for a meeting that already happened.
MineTwin answers the questions operators haven't thought to ask yet. It looks across every domain - maintenance, production, safety, workforce, costs - and surfaces the correlations that don't show up in any single report.
When a VP of Operations asks "where is this mine losing money nobody has flagged yet" - the system delivers three specific answers, with the data to back them, in under ten seconds. That's not a dashboard. That's an analyst who knows the mine as well as the best people on site - and never goes home.
This is what every mine site a customer operates could have. The data already exists in their systems. MineTwin builds the intelligence layer on top of it.
Proven data ingestion patterns: Fleet management (MineStar, Wenco, Komatsu), CMMS (SAP, Maximo), environmental monitoring, HR, and financials - all flowing into a Databricks lakehouse. The architecture is organized in three layers: reference data (equipment registers, employee rosters, PM plans), operational data (production, telemetry, work orders, incidents), and pre-computed views (23 common questions already answered, covering 80% of queries in under 10 seconds). This architecture is repeatable - every new customer deployment follows the same pattern.
MineTwin connects to the infrastructure you already have. Your data stays in your environment. Your team doesn't change how they work - they just gain a layer of intelligence on top of what already exists.
We connect to your fleet management system (MineStar, Wenco, Komatsu), your CMMS (SAP, Maximo), environmental monitoring, HR, and financials. Read-only integration - nothing is replaced, nothing is disrupted. The data flows into a Databricks lakehouse on your Azure tenant. Your IT team approves once.
We build the 3D twin of your site using satellite terrain, your engineering drawings, and our asset library - 68 mine assets ready to go, from haul trucks to flotation cells. Your site-specific equipment maps to our digital twin schema. Every relationship is mapped.
We set up your role profiles, safety targets, KPI definitions, and environmental limits. The AI connects to your data and documents - no model training required. 9 role profiles covering 150+ job titles are mapped to your org structure automatically.
We write the narration for your site with your team. Each step references your actual equipment, your actual procedures, your actual numbers. Four durations - 5, 10, 20, and 60 minutes. The platform is the same. The data and content are yours. 4–8 weeks from data access to live environment.
MineTwin connects to your existing systems, builds the twin, and is live before your next rotation. Here's exactly what that looks like operationally.
Fleet management (MineStar, Wenco, Komatsu), CMMS (SAP, Maximo), environmental monitoring, HR, financials, and your document store - SOPs, safety plans, operating manuals. Read-only integration into a Databricks lakehouse on your Azure tenant. Your IT team approves once. Nothing is modified in your source systems.
We build the 3D twin from satellite terrain, your engineering drawings, and our asset library - 68 mine assets across 20 categories, ready to go. Every asset placed, every relationship mapped, your operational history loaded. Your data stays inside your security boundary.
Role profiles mapped to your org structure - 9 profiles covering 150+ job titles, from GM to apprentice fitter. Safety targets, KPI definitions, environmental limits. The AI connects to your data and 22 embedded documents. No model training - configuration, not customization.
Orientation scripts co-written with your team - each step references your actual equipment, procedures, and numbers. Four durations (5, 10, 20, 60 minutes), scripted narration with live Q&A, completion logged automatically. The system goes live before your next crew rotation. 4–8 weeks from data access to live environment.
The direct labor savings are real and measurable. But the strategic return - your best people back on the job, institutional knowledge that stays when people leave - is what compounds over time.
Direct labor is the cost that shows up on a budget line. It's real, it's measurable, and it's significant. But the cost you can't invoice is the one that costs you most.
Every rotation, people from HR, health & safety, and operations step out of their roles to run inductions - four to six hours, every time, for every new arrival. The same standardized content, delivered by whoever is available. Multiply that across a year and you've built an invisible second job on top of three departments you can least afford to distract.
MineTwin gives that time back. And six months after deployment, you have something more valuable than the time savings: a system that knows your mine as well as those people do - and never leaves.
The induction is the entry point. The institutional memory is the return.
Conservative model - direct labor only. No retention assumptions, no safety incident reductions, no time-to-productivity gains. Just the instructor hours you stop spending.
The $749K figure is what you can put in a spreadsheet and defend. It's based on 2,350 induction events per year - new hires, rotating contractors, site visitors - at 6 hours per event, 1.5 staff per event, $85/hr loaded cost, saving 2.5 hours per event.
That's a conservative number. It only counts the direct time spent delivering inductions. It doesn't count the full cost across HR, health & safety, and operations - three departments all contributing staff to deliver the same standardized content, every rotation, instead of doing their actual jobs.
Add those in and the real cost per site is higher. The savings are higher. The payback is faster.
The model below uses the conservative inputs. Adjust them to your site and the number becomes yours to own.
Tier 1 - induction and orientation - is the beachhead. The same platform, the same twin, the same data architecture supports Tier 2 operational agents and Tier 3 enterprise intelligence. Each tier is incrementally cheaper to build and incrementally more valuable per site.
The Tier 1 license is priced to be an easy yes for a GM - payback in under 90 days, no new infrastructure, no rip-and-replace. It's designed to get the twin deployed, not to be the end state.
Once the twin is running on a site, Tier 2 unlocks: haulage optimisation, predictive maintenance, concentrator throughput, fuel efficiency. Same platform. Live telemetry layer added. Value per site increases by a multiple, build cost is incremental because the twin and data architecture already exist.
Tier 3 sells to the C-suite: executive briefing, cross-site benchmarking, portfolio optimisation. At that point MineTwin is embedded in how the company makes decisions - not just how it onboards people.
The moat compounds with every deployment. Site-specific data models, ERP integrations tuned to each mine's schema, role-trained AI that learns the operation over months. Every site becomes harder to displace - not because of lock-in, but because the system genuinely knows the mine better than any standing-start alternative ever could.
Real-time, data-backed answers grounded in actual mine systems became production-grade reliable in 2025. Not impressive demos - operational reliability. The technology to build MineTwin exists now in a way it didn't two years ago.
Trimble, Hexagon, and Bentley sell asset registers and maps. None of them know who the workers are. The category of intelligent, role-aware mine operations knowledge is completely open - and the first platform to build a data moat wins it.
Mining is losing experienced workers faster than it's replacing them. FIFO operations re-onboard the same positions every rotation. Safety regulators are tightening. C-suites need a solution that actually scales. MineTwin is it.
Role-aware 3D induction and ongoing AI guide. Connects to existing mine data. $100K+ per site per year. Deployed in weeks. Payback in under 90 days.
Haulage optimisation. Predictive maintenance. Concentrator throughput. Fuel efficiency. Dispatch advisory. Same platform, live telemetry layer added. Built on the twin already deployed - incremental build time, step-change in value per site.
Executive briefing. Cross-site benchmarking. Portfolio optimisation. Board and investor intelligence. Sells to the C-suite. Each tier costs less to build than the last - the twin and data architecture are already there.
300–1,000 employees · Remote or FIFO operations · 30%+ annual turnover · High regulatory scrutiny · Multiple sites under common management
800+ mines globally meet this profile. We don't need them all. We need 10 - and we already know how to get them.
MineZero is our proving ground - a synthetic mine that demonstrates the platform's data ingestion patterns, AI workflows, and 3D environment at production scale. The same platform deploys to a customer site in weeks.
Committed at founding terms - pre-seed raise underway via YC SAFE at $10M cap
Operational records across 68 Databricks tables - 4 years of production, maintenance, safety, workforce, and financial data
Databricks tables across fleet, CMMS, environmental, HR, and financials - SAP, Maximo, MineStar ingestion patterns proven
Equipment assets modeled with full ERP ID mapping and 4-year maintenance history in the MineZero proving ground
Synthetic staff profiles structured to mirror a real mine HR schema - org hierarchy, FIFO rotations, certifications
Pre-computed views covering 80% of common operational questions - answers in under 10 seconds
Role profiles covering 150+ job titles - from GM to apprentice fitter, with data access controlled at the AI level
From data access to live deployment at a customer site - the platform is proven, only the data connection and site content change
Customer discovery interviews underway - pilot site conversations with mine operators in progress
Preston built the system MineTwin replaces - inside one of the world's largest mining companies, at enterprise scale, with NVIDIA backing. Ludvig built the 3D engine it runs on. No other founding team has both mining operations depth and proprietary spatial technology.
Founder & CEO of MineTwin - building the platform that connects data systems, people, and AI through a shared spatial context. That shared context is what makes everything else work.
Mine engineer first - underground, open pit, Arctic. Over a decade in mining operations and technology leadership, including building a global-scale digital twin program at Teck. The problem was always the same: critical knowledge trapped in people's heads with no shared reference for how the operation works.
Now I use AI to make that vision real - faster and better than was ever possible before.
Creator of TurboTwin - the proprietary 3D spatial engine that powers MineTwin. Purpose-built for industrial-scale real-time environments. This is not game technology adapted for mining. It was designed for operations like this.
Arvikasoft CEO - runs a software studio specializing in real-time 3D and industrial data systems. The development partner behind MineTwin's technical build.
Unreal Engine practitioner since 2014. A civil engineer who builds 3D engines - the exact combination this platform requires.
Preston built the system MineTwin replaces - inside one of the world's largest mining companies, at enterprise scale, with NVIDIA backing. Ludvig built the 3D engine it runs on. No other founding team has both mining operations depth and proprietary spatial technology.
Founder & CEO of MineTwin - building the platform that connects data systems, people, and AI through a shared spatial context. That shared context is what makes everything else work.
Mine engineer first - underground, open pit, Arctic. Over a decade in mining operations and technology leadership, including building a global-scale digital twin program at Teck. The problem was always the same: critical knowledge trapped in people's heads with no shared reference for how the operation works.
Now I use AI to make that vision real - faster and better than was ever possible before.
Creator of TurboTwin - the proprietary 3D spatial engine that powers MineTwin. Purpose-built for industrial-scale real-time environments. This is not game technology adapted for mining. It was designed for operations like this.
Arvikasoft CEO - runs a software studio specializing in real-time 3D and industrial data systems. The development partner behind MineTwin's technical build.
Unreal Engine practitioner since 2014. A civil engineer who builds 3D engines - the exact combination this platform requires.
A 30-day pilot on your site. No new infrastructure. No disruption to existing systems. Your data, your mine, your people - answered by an AI that already understands the operation.
Your HR, H&S, and operations people are spending four hours per induction event in a classroom instead of doing their jobs. Your institutional knowledge walks out the door every time someone retires or rotates. Your analytics team takes a week to produce what this answers in seconds.
You don't need to take our word for it. Run it on your site. Ask it questions about your operation. Let your team decide whether it earns its place.
We'll connect to your existing data, build a digital twin of your site, and deploy a working AI guide - in weeks, not months. At the end of 30 days, you'll know whether this belongs in your operation.
No purchase commitment. No new infrastructure. We work with what you have.
Start the Conversation →Contact
Preston Miller, P.Eng., MBA · preston.miller@minetwin.ai
Vector Foundries Corporation · minetwin.ai
30 days. We connect to your existing data. Your team asks it questions about your mine. You decide if it earns its place.
You know the problem. Your best H&S lead runs induction and it's excellent. Whoever could be spared from HR or operations runs it and it shows. Same slides. Different delivery. You can't fix it with headcount you don't have.
MineTwin is the headcount. Consistent, role-aware, available every rotation - without pulling anyone off the job.
We'll work with your IT team to set up a read-only connection to your existing systems. No modifications to your source data. No new infrastructure to procure. The twin is built from what you already have.
At the end of the pilot, your team will have used it. That's the business case.
Start the Conversation →Contact
Preston Miller, P.Eng., MBA · preston.miller@minetwin.ai
Vector Foundries Corporation · minetwin.ai
We're raising $500K to get from working product to first paying customers. Every dollar maps to a specific milestone. The playbook is written.
If you're an investor, you're looking at a capital-efficient SaaS business in a category with no incumbent, being built by the only founders who have done this from both sides. The beachhead market is well-defined, the unit economics are strong, and the expansion path into operational intelligence multiplies the value per site without starting over.
We don't need to convince mining operators that this problem is real. They're living it. We need to show them the solution - and we've built it.
We offer a 30-day pilot - connect to the operator's existing data, build the twin, let their team ask it questions about their operation. At the end of 30 days, the mine decides whether their analysts should keep spending a week on what this answers in seconds.
No purchase commitment required from the operator. No new infrastructure. We work with what they have.
Start the Conversation →Investor contact
Preston Miller · preston.miller@minetwin.ai
Vector Foundries Corporation · Delaware C-Corp · Standard YC SAFE
This page is for informational purposes only and does not constitute an offer to sell or a solicitation of an offer to buy any securities. Any offering of securities will be made solely to accredited investors pursuant to definitive offering documents and applicable exemptions from registration under the Securities Act of 1933. Figures, projections, and forward-looking statements reflect current expectations, involve known and unknown risks, and are not guarantees of future performance. Nothing on this page constitutes investment, legal, accounting, or tax advice. Prospective investors should consult their own advisors before making any investment decision.