AquaTwin Agent is an AI agent for governance-grade water accounting in mining operations. Drag-and-drop visual modeling, contextualized document intelligence, weighted state fusion, and auditable water balance in Excel, aligned with ICMM WAF, GISTM, and TSM Water Stewardship.
Embedded live on desktop, it opens pre-populated with Mina Data Riders, a seed iron ore operation. On mobile, tap the button below to open the full experience in a dedicated window.
Talk to the AI, ask anything, get answers
This is a real conversational agent, not a static demo. Ask about any metric, request explanations of results, dig into the data, or simply type "How does this work?", it responds in real time.
Open AquaTwin Agent in a new tab
Open AquaTwin Agent
Best experienced in a dedicated window on mobile.
The output is a semantic digital twin of the water system, designed for governance-grade water accounting.
Visual modeling
Drag-and-drop editor for the operation's water network: canonical asset types, directed water flows, and WAF zones (Mine, Beneficiation, Dams, Environment, Treatment).
Deterministic parameterization
Each asset carries a fixed schema of properties by type (dam capacity, specific plant consumption, treatment efficiency...) and the list of measurements it must report over time.
Contextualized document intelligence
Ingestion of PDF, Excel, Word, and PowerPoint. Each extracted observation is linked to an asset or flow, period, canonical unit, and nature (measured, simulated, estimated), with traceability down to the page, cell, or slide.
Weighted state fusion
Kalman filter-inspired fusion over time: each observation updates the state with a weight derived from the declared reliability of the evidence. A high-reliability measurement outweighs multiple low-confidence estimates.
Water balance & Excel export
Closes the water balance on demand, monthly, or at your required cadence. Exports a multi-tab Excel workbook with an accuracy statement per variable, a navigable evidence package, and a plate code for version integrity.
Governance-grade output
The Excel artifact is designed to look like a controlled engineering sheet, auditable, signed, and dated, not a dashboard export. Ready for WAF reporting and GISTM/TSM evidence packages.
Environmental engineers
Water governance specialists at major mining companies, iron ore archetype, already familiar with ICMM WAF, GISTM, and TSM.
Operations managers
Responsible for closing the complex's water balance, requiring traceable figures and a defensible engineering sheet, not a dashboard.
Sustainability
Responsible for water reporting across WAF / GISTM / TSM cycles, looking for evidence-first AI, not generative guesswork.
The MVP opens pre-populated with Mina Data Riders, a Brazilian seed iron ore operation, and a curated document corpus pre-processed into validated observations. Public, no login required, users enter a live system.
Seed iron ore operation · curated document corpus · on-demand water balance closing · email preview · downloadable Excel.
Scheduled dispatch, richer state fusion parameters, SCADA/historian integration, internal plant breakdown, and expanded coverage of WAF quality parameters, shaped alongside early partners.
Run AquaTwin Agent in a new tab with Mina Data Riders already modeled, an active document pipeline, and a closed balance ready for download in Excel. No sign-up, no friction.
The editorial and engineering stance behind the digital twin, explained in long form.
How governance maturity translates into water stewardship under TSM.
The metrics that matter when closing the monthly balance.
How AI reading technical documents turns files into traceable, auditable evidence.
Where tailings governance meets water balance, and why they cannot be separated.
AquaTwin Agent is the engine. These case studies show the water governance lineage behind it: methodology, KPIs, instrumentation, and the agent in operation.
AI agent for water balance & governance
The architectural predecessor of AquaTwin: curated knowledge base, tight system prompt, curatorial loop.
Water management pilot at a mining complex
Diagnostics, KPIs, instrumentation, and AI agent in operation: the full stack AquaTwin was designed to serve.
Data reliability methodology
Auditable accuracy per variable, the methodology compiled into the AquaTwin fusion layer.
WAF-aligned water KPI tree
The indicator hierarchy behind AquaTwin's accuracy statement, traceability by design.
See it in action
Real agent screens, with fictional data.
