GeonatIQ’s Exploration & Resource Signal Agent is designed to monitor, detect, and interpret early technical and data-driven signals related to resource discovery, reserve revisions, and changes in resource quality. The system focuses on upstream, mining, and subsurface storage-relevant signals that often emerge first in technical disclosures and data releases before being reflected in market or company-level announcements. The framework operates across energy, mining, critical minerals, and subsurface storage sectors and can be configured to track specific commodities, basins, resource types, and technical themes. It is grounded in real-world exploration, appraisal, and reserve reporting workflows.
Inputs: The agent ingests data from technical papers, government and geological survey releases, reserve and resource reports, academic publications, conference proceedings, and industry announcements. Inputs are tagged by commodity, resource type, basin or geological setting, reporting standard, data quality, and development stage. Historical exploration and reserve data is used to benchmark significance and reliability.
AI Modelling Approach: The system continuously scans for technical signals including new discovery indicators, reserve upgrades or downgrades, changes in grade, permeability, storage capacity, or recovery factors. Signals are classified by credibility, materiality, and relevance to specific resource strategies. The model contextualises findings against regional geology, historical outcomes, and comparable assets, updating assessments as additional data is released.
Outputs & Use Cases: Clients receive structured intelligence and early-warning alerts on exploration and resource-related developments aligned with their strategic focus. Outputs include concise signal summaries, technical context, and relevance flags. The agent supports upstream strategy, exploration planning, portfolio optimisation, and subsurface storage evaluation by surfacing material resource signals ahead of broader market recognition.
GeonatIQ’s Well Monitoring Intelligence Agent is designed to act as a continuous decision-support layer for operational geologists and subsurface teams, providing real-time monitoring, anomaly detection, and contextual checks across active wells. The system functions as a technical back-up brain, helping reduce cognitive load during operations by continuously watching for deviations, inconsistencies, and emerging risks that may be missed in fast-moving operational environments. The framework operates across drilling, completion, production, and subsurface storage wells and can be configured by the user to monitor specific wells, formations, parameters, and operational objectives. It is grounded in real-world well operations and geoscience workflows.
Inputs: The agent ingests real-time and near-real-time operational data including drilling parameters, mud logs, wireline and LWD data, pressure and temperature readings, integrity metrics, and geological interpretations. User-defined monitoring objectives and thresholds are incorporated, alongside historical well data and analogue cases for context.
AI Modelling Approach: The system continuously analyses incoming data to detect anomalies, geological inconsistencies, and integrity risks relative to expected stratigraphy, petrophysical models, and operational plans. Signals are classified by severity, confidence, and operational relevance. The model provides contextual guidance by comparing current conditions with historical outcomes and known failure modes, supporting rapid, informed decision-making.
Outputs & Use Cases: Operational teams receive real-time alerts, contextual insights, and decision-support prompts during drilling and well operations. Outputs include anomaly flags, consistency checks, integrity warnings, and suggested lines of inquiry for ops calls. The agent supports safer, faster, and more confident operational decisions by reducing cognitive burden and acting as a persistent technical co-pilot.
GeonatIQ’s M&A Pre-Signal Intelligence Agent is designed to detect early indicators of acquisition or divestment intent across energy, commodities, natural resources, and climate-related sectors. The system focuses on pre-transaction signals that emerge before formal deal processes are launched, helping clients anticipate strategic moves rather than reacting once transactions become public. The framework operates across upstream and downstream energy, mining, infrastructure, carbon, and climate-adjacent asset classes and can be configured to track specific companies, assets, regions, and strategic themes. It is grounded in real-world corporate strategy, capital allocation, and transaction dynamics.
Inputs: The agent ingests data from company disclosures, capital allocation statements, portfolio rationalisation announcements, executive commentary, board changes, advisor appointments, regulatory filings, and selected industry and financial publications. Inputs are tagged by company, asset type, commodity exposure, geography, balance sheet context, and strategic intent indicators. Historical transaction data is used to establish baseline pre-deal patterns.
AI Modelling Approach: The system continuously scans for behavioural and strategic signals associated with M&A intent, including shifts in capital spending, asset write-downs, strategic reviews, partnership exits, and management language changes. Signals are classified by strength, credibility, and likely transaction pathway. The model learns from prior transactions to refine signal weighting and improve early detection accuracy.Outputs & Use Cases: Clients receive early-warning intelligence on potential acquisitions or divestments relevant to their strategic focus. Outputs include pre-signal alerts, contextual summaries, and confidence assessments. The agent supports corporate strategy, origination, due diligence preparation, and competitive intelligence by surfacing likely M&A activity ahead of formal deal announcements.
GeonatIQ’s ESG Risk & Opportunity Agent is designed to monitor, interpret, and prioritise environmental, social, and governance signals across energy, commodities, natural resources, and climate-related assets. The system focuses on identifying material ESG risks and positive developments early, translating complex and often fragmented ESG information into decision-grade intelligence rather than generic scoring or box-ticking assessments. The framework operates across upstream energy, mining, infrastructure, carbon, and climate assets and can be configured to track specific companies, projects, jurisdictions, and ESG themes. It is grounded in real-world regulatory, operational, and stakeholder dynamics.
Inputs: The agent ingests data from regulatory actions, permitting decisions, litigation and enforcement notices, NGO and civil society reports, community engagement disclosures, accident and incident reports, environmental monitoring data, labour and supply chain disclosures, and corporate sustainability announcements. Inputs are tagged by ESG pillar (E, S, G), theme (for example emissions, biodiversity, water, labour, governance), jurisdiction, asset or project linkage, and potential financial or operational exposure. Historical ESG events are used to establish baseline risk and outcome patterns.
AI Modelling Approach: The system continuously scans for ESG-related signals and classifies them by severity, credibility of source, and time horizon. Risks and opportunities are assessed based on their potential impact on permitting, timelines, capex, opex, insurance, cost of capital, and reputational exposure. The model differentiates between transient noise and structurally material issues and adapts dynamically as situations escalate, resolve, or improve.
Outputs & Use Cases: Clients receive early-warning alerts and structured summaries of ESG risks and positive developments aligned with their assets, investments, or strategic priorities. Outputs include materiality flags, impact context, and escalation indicators. The agent supports risk management, investment analysis, stakeholder engagement, and strategic decision-making by surfacing ESG signals early and in a form that can be acted upon.
GeonatIQ’s Capital Allocation & Capex Signal Agent is designed to monitor and interpret shifts in corporate capital allocation across energy, mining, natural resources, and climate infrastructure sectors. The system focuses on early signals related to changes in capex plans, project sanctioning, deferrals, expansions, and divestments, translating corporate disclosures into forward-looking investment intelligence rather than backward-looking reporting. The framework operates across upstream, midstream, downstream, mining, power, carbon, and climate infrastructure assets and can be configured to track specific companies, commodities, regions, and project types. It is grounded in real-world capital planning and corporate finance dynamics.
Inputs: The agent ingests data from corporate announcements, earnings calls and transcripts, investor presentations, regulatory filings, capital markets disclosures, and selected industry publications. Inputs are tagged by company, asset or project, commodity exposure, geography, capex category, timeline, and stated strategic rationale. Historical capex patterns and project outcomes are used to establish baseline behaviour.
AI Modelling Approach: The system continuously scans disclosures and management commentary for signals indicating changes in capital allocation, including revised capex guidance, project approvals or cancellations, portfolio rationalisation, and funding reallocations. Signals are classified by materiality, confidence, and likely execution risk. The model contextualises announcements against balance sheet capacity, prior commitments, and peer behaviour to assess credibility and follow-through likelihood.
Outputs & Use Cases: Clients receive structured intelligence and early-warning alerts on capex and capital allocation shifts relevant to their strategic or investment focus. Outputs include signal summaries, project-level context, and confidence assessments. The agent supports investment analysis, strategic planning, portfolio monitoring, and competitive intelligence by surfacing capital allocation changes as they emerge.
GeonatIQ’s Regulatory News Round-Up Agent is designed to monitor, filter, and synthesise regulatory developments across selected interest areas, translating high-volume policy and regulatory updates into concise, decision-relevant intelligence. The system focuses on what has changed, what is coming, and what matters for a defined regulatory or commercial focus, rather than reproducing raw announcements. The framework operates across commodities, energy, mining, climate, and infrastructure sectors and can be configured by the user to track specific jurisdictions, regulators, policy themes, and asset types. It is grounded in real-world regulatory processes and enforcement dynamics.
Inputs: The agent ingests data from government departments, regulators, ministries, legislative bodies, consultation papers, enforcement notices, and selected legal and industry publications. Inputs are tagged by jurisdiction, regulator, policy area, commodity or sector relevance, implementation timeline, and potential commercial impact. Historical regulatory actions are used to provide context and trend detection.
AI Modelling Approach: The system continuously scans regulatory sources and classifies updates by materiality, relevance, and time horizon. Items are prioritised based on their likely impact on operations, permitting, compliance, capital allocation, or market structure. The model adapts as consultations progress into formal rules and enforcement, refining relevance as regulatory clarity increases.Outputs & Use Cases: Clients receive a scheduled regulatory round-up delivered daily, or at a chosen frequency, summarising the most relevant regulatory developments aligned with their specified interest area. Outputs include concise summaries, impact flags, and timeline indicators. The agent supports compliance monitoring, strategic planning, and risk management by delivering timely regulatory intelligence directly to the inbox.
GeonatIQ’s Supply Chain & Critical Inputs Risk Agent is designed to monitor, detect, and contextualise risks across commodity supply chains and critical input dependencies that could disrupt operations, project timelines, or financial performance. The system focuses on early warning signals rather than realised disruptions, identifying stress points in materials, logistics, processing, and supplier networks before they translate into operational or commercial impact. The framework operates across energy, mining, critical minerals, and industrial supply chains and can be configured to track specific commodities, inputs, suppliers, regions, and transport routes. It is grounded in real-world supply chain dynamics and industrial risk management.
Inputs: The agent ingests data from company disclosures, supplier announcements, trade and customs data, logistics and shipping indicators, government and regulator updates, industry publications, and selected alternative data sources. Inputs are tagged by commodity, critical input type, supplier, geography, transport mode, and dependency level. Historical disruption events are incorporated to provide context and baseline risk patterns.
AI Modelling Approach: The system continuously scans for early indicators of supply chain stress including production outages, export restrictions, permitting delays, labour actions, logistics bottlenecks, sanctions, and force majeure events. Signals are classified by severity, credibility, and likely duration. The model prioritises risks based on criticality of the input, availability of substitutes, concentration risk, and downstream exposure, updating assessments as conditions evolve.
Outputs & Use Cases: Clients receive early-warning alerts and structured intelligence on emerging supply chain and critical input risks aligned to their operations, investments, or trading exposures. Outputs include risk summaries, dependency mapping, and escalation indicators. The agent supports operational planning, procurement strategy, risk management, and investment decision-making by surfacing supply chain vulnerabilities before they materially affect outcomes.
GeonatIQ’s Litigation & Dispute Early-Warning Agent is designed to monitor, detect, and contextualise emerging legal disputes, regulatory actions, and litigation risks across commodities, energy, mining, and infrastructure sectors. The system focuses on early signals rather than headline outcomes, identifying where legal, contractual, or regulatory friction is building before it becomes fully priced into assets, valuations, or counterparties. The framework operates across jurisdictions and legal regimes and can be configured to track specific companies, projects, commodities, contract types, and regulatory bodies. It is grounded in real-world dispute dynamics and enforcement processes.
Inputs: The agent ingests data from court filings, arbitration registers, regulatory enforcement notices, tribunal updates, company disclosures, government agencies, and selected legal and industry publications. Signals are tagged by dispute type (contractual, regulatory, environmental, labour, IP, permitting), jurisdiction, parties involved, project or asset linkage, and potential financial or operational exposure. Historical dispute data is used to establish baseline risk patterns.
AI Modelling Approach: The system continuously scans for early indicators of disputes, including procedural filings, regulator correspondence, licence challenges, community objections, and enforcement escalations. Events are classified by severity, credibility, and likely trajectory. The model prioritises signals based on potential impact to operations, timelines, capital structure, or reputational risk, and updates assessments as disputes progress.
Outputs & Use Cases: Clients receive early-warning alerts and structured summaries of emerging litigation and dispute risks aligned to their assets, investments, or counterparties. Outputs include risk categorisation, timeline tracking, and escalation indicators. The agent supports risk management, due diligence, portfolio monitoring, and strategic decision-making by surfacing legal risks early, before they materially affect outcomes.
GeonatIQ’s Latest Discoveries Intelligence Agent is designed to monitor and synthesise new discovery announcements across selected commodities, identifying material finds as they are disclosed and translating them into structured, decision-ready intelligence. The system focuses on verified discovery events rather than speculation, capturing the technical and commercial context required to assess relevance and impact. The framework operates across energy, mining, critical minerals, and adjacent natural resource sectors, and can be configured by the user to track specific commodities, regions, and discovery types. It is grounded in real-world exploration and appraisal workflows.
Inputs: The agent ingests data from company announcements, regulatory filings, stock exchange disclosures, technical reports, industry publications, and selected government and geological survey sources. Each discovery is tagged by commodity, location, geological setting, reported size or resource estimate, development stage, partners, operatorship, and jurisdictional context. Historical discovery data is used to benchmark scale and significance.
AI Modelling Approach: The system continuously scans for new discovery disclosures and classifies them by materiality, credibility of source, and strategic relevance. Discoveries are contextualised against regional geology, historical success rates, and comparable finds to assess relative significance. The model adapts as additional technical data is released, refining assessments as discoveries move from announcement to appraisal.
Outputs & Use Cases: Clients receive timely intelligence on new discoveries aligned with their commodity focus, including structured summaries covering location, size, partners, and development context. Outputs can be delivered as real-time alerts or scheduled briefings. The agent supports exploration strategy, competitive intelligence, investment analysis, and partnership development by surfacing and contextualising new discovery events as they occur.
GeonatIQ’s Human Capital Intelligence Agent is designed to monitor employment movements and talent signals across LinkedIn and selected industry-specific platforms chosen by the user. The system focuses on detecting meaningful role changes, senior appointments, departures, and team expansions that align directly with defined recruitment priorities. It is built to surface actionable human capital intelligence rather than raw job postings or undifferentiated profile updates. The framework operates across energy, commodities, mining, climate, and technology-adjacent sectors, and can be configured by the user to track specific role categories, seniority levels, skills, geographies, and target organisations. It is grounded in real-world recruitment workflows and market behaviour.
Inputs: The agent ingests data from LinkedIn activity signals and user-selected industry sources such as trade publications, professional associations, conference speaker lists, and sector-specific hiring platforms. Individuals and organisations are tagged by role, function, seniority, sector relevance, skills, and geography. Historical role transitions and firm-level hiring patterns are used to provide context and trend awareness.
AI Modelling Approach: The system continuously monitors profile changes and industry announcements and classifies employment events by relevance and priority against the user’s recruitment criteria. Signals are weighted based on seniority, skill scarcity, employer type, and timing. The model learns from engagement and hiring outcomes to refine signal quality, reduce noise, and prioritise high-conviction talent movements.
Outputs & Use Cases: Clients receive scheduled intelligence updates, daily or at a chosen frequency, highlighting the most relevant talent movements across LinkedIn and selected industry sources. Outputs include concise summaries of key hires, departures, emerging candidate pools, and organisational shifts. The agent supports proactive recruitment, competitor monitoring, and strategic workforce planning by delivering timely, targeted human capital insights.
GeonatIQ’s Bid Rounds Intelligence Agent is designed to monitor, interpret, and prioritise new government-led bid rounds across commodities by tracking licensing announcements, regulatory timelines, and policy signals. The system focuses on identifying where and when new acreage, permits, or concessions are becoming available and aligning these opportunities with specific commodity interests and strategic criteria. It is built to convert fragmented regulatory information into early, actionable opportunity intelligence rather than reactive deal tracking. The framework operates across energy, mining, carbon storage, hydrogen, and adjacent natural resource sectors, and is designed to scale across jurisdictions with differing regulatory regimes and disclosure standards. It is grounded in real-world licensing processes and government tender structures.
Inputs: The agent ingests data from government agencies, regulators, ministries, and licensing authorities, including formal bid round announcements, draft consultations, regulatory calendars, policy updates, and legislative changes. This is complemented by jurisdictional data on fiscal terms, environmental requirements, acreage characteristics, and historical award outcomes. Each bid round is tagged by commodity, jurisdiction, timing, entry requirements, technical criteria, and strategic relevance.
AI Modelling Approach: The system continuously scans for early indicators of upcoming bid rounds, including consultations, policy shifts, and pre-announcement signals. Identified rounds are mapped against commodity-specific interest profiles and strategic filters such as resource type, regulatory risk, timeline, and capital intensity. The model prioritises opportunities based on alignment, readiness, and relative attractiveness, and adapts dynamically as bid terms and timelines evolve.
Outputs & Use Cases: Clients receive forward-looking intelligence on upcoming and open bid rounds, ranked by relevance to their commodity focus and strategic objectives. Outputs include opportunity summaries, timeline tracking, and alerts as rounds move from consultation to application and award. The agent supports early positioning, partnership formation, and capital planning by surfacing bid opportunities before they become widely competitive.
GeonatIQ’s Capital Raising Agent is built to continuously identify, qualify, and prioritise investors across multiple commodity and real asset themes by monitoring live mandate signals, disclosed investment interests, and behavioural indicators. It is not a static CRM or keyword-based search tool. The system is designed to understand investor intent, timing, and mandate alignment in real time, translating fragmented public and proprietary signals into actionable capital raising intelligence. Context and mandate fit sit at the core of the model. The framework operates across energy, mining, carbon, hydrogen, infrastructure, and adjacent real asset strategies, and is designed to scale as new commodities, structures, and investor behaviours emerge. It is grounded in AIP’s proprietary investor intelligence workflows and active capital raising experience.
Inputs: The agent ingests structured and unstructured investor data including fund mandates, allocation statements, recent fundraises, LP disclosures, public filings, press releases, conference participation, and real-time updates on stated investment interests. This is combined with AIP’s proprietary investor database, historical engagement data, and partner-sourced intelligence. Each investor is tagged by asset class, commodity exposure, investment stage, geography, cheque size, timing indicators, and strategic focus.
AI Modelling Approach: The system continuously scans for mandate changes and live investment signals and maps them against specific capital raising requirements. Investors are dynamically scored based on relevance, recency of interest, historical behaviour, and inferred likelihood of engagement. The model learns from prior successful capital raises to refine targeting logic, suppress false positives, and surface high-conviction prospects as conditions change.
Outputs & Use Cases: Clients receive ranked, decision-grade investor target lists tailored to their specific capital raise, along with real-time alerts as new aligned investors emerge or existing mandates shift. The agent supports fundraises, strategic placements, project-level financings, and partnership development by focusing outreach on investors who are both mandate-aligned and actively deploying capital.
GeonatIQ’s AI Porosity & Storage Quality Detection Model is designed to identify subsurface porosity and reservoir-quality signatures relevant to CO₂ and hydrogen storage. The model analyses seismic data in high-dimensional attribute space to detect patterns associated with porous, laterally continuous, and potentially well-connected reservoir units. It is intended as a screening and decision-support tool, used in combination with expert geoscientist interpretation to help pinpoint and prioritise onshore and offshore storage opportunities, rather than as a standalone or foolproof solution.
Inputs: The model ingests post-stack seismic volumes together with a focused set of seismic attributes that are known to be sensitive to porosity, lithology, and reservoir quality. These typically include amplitude and energy attributes, frequency and bandwidth measures, attenuation- and Q-related attributes, phase and continuity indicators, and selected structural attributes that provide depositional and trapping context. Client seismic datasets, attribute selections, and optional calibration data (e.g. known reservoirs or analogue fields) can be incorporated.
AI Modelling Approach: The model is trained on the attribute character of porous and reservoir-quality intervals expressed in high-dimensional attribute space. Rather than relying on single attributes or simple thresholds, it learns multivariate patterns that collectively reflect porosity, grain framework, attenuation behaviour, and continuity at reservoir scale. Machine learning is used to generalise these patterns across large seismic volumes, enabling systematic scanning for intervals that exhibit similar porosity-related signatures. The approach is explicitly probabilistic and designed to work alongside geoscientist expertise, allowing geological context, seal integrity, and storage suitability to be assessed in parallel.
Outputs, Validation & Performance: Clients receive probabilistic porosity and storage-quality indicator volumes, ranked zones of interest, and spatially coherent reservoir candidates suitable for further evaluation for CO₂ or H₂ storage. Outputs can be integrated directly into standard interpretation platforms and reviewed alongside structural, stratigraphic, and seal analyses. Validation is carried out through comparison with known porous intervals, analogue fields, and spatial generalisation tests across different areas. Results are delivered as seismic-compatible outputs or through a GeonatIQ AI agent, supporting rapid regional screening, portfolio prioritisation, and informed expert-led decision-making for subsurface storage projects.
GeonatIQ’s AI DHI Detection Model is designed to identify and prioritise potential Direct Hydrocarbon Indicators within seismic data by analysing their characteristic attribute signatures in high-dimensional space. The model supports exploration teams in screening for hydrocarbon-related responses, highlighting areas of elevated prospectivity, and focusing detailed interpretation effort where it matters most. Use cases include early-stage prospect ranking, de-risking leads and prospects, and supporting integration with inversion, AVO, and rock physics workflows.
Inputs: The model ingests post-stack seismic volumes and a targeted set of seismic attributes known to be sensitive to DHI expression. These include amplitude-based attributes (e.g. RMS amplitude, envelope, sweetness), frequency and bandwidth measures, phase and polarity attributes, and selected structural context attributes. The model can be trained on client data and on examples of known or interpreted DHIs, capturing their attribute behaviour without relying on explicit spatial rules or hard thresholds.
AI Modelling Approach: The DHI model is trained on the attribute character of known Direct Hydrocarbon Indicators represented in high-dimensional attribute space. Rather than searching for a single amplitude response, the model learns multivariate patterns and associations that collectively describe DHI behaviour, including subtle combinations of amplitude, frequency, and phase responses. Machine learning is used to identify and generalise these patterns across the seismic volume, enabling the model to scan large datasets and highlight areas with similar attribute signatures. The approach is intentionally probabilistic and data-driven, recognising that DHI expression is context-dependent and not foolproof.
Outputs, Validation & Performance: Clients receive probabilistic DHI likelihood volumes, ranked anomaly clusters, and spatially coherent indicators that can be reviewed alongside seismic, inversion, and AVO results. Outputs are designed to guide attention rather than replace geological judgement, helping teams to quickly narrow down candidate areas for further analysis. Validation is performed through comparison with known hydrocarbon occurrences, internal consistency checks, and spatial generalisation tests. Results are delivered as seismic-ready outputs or through a GeonatIQ AI agent, supporting rapid screening, iterative refinement, and informed decision-making while explicitly acknowledging uncertainty and risk.
GeonatIQ’s AI Oil & Gas Exploration Model applies advanced machine learning to large-scale seismic and subsurface datasets to accelerate prospect screening, facies mapping, and geological interpretation. The model is designed to support exploration teams in identifying stratigraphic and structural patterns, reducing manual interpretation effort, and improving consistency across regional and field-scale studies. Use cases include early-stage exploration screening, play fairway analysis, reservoir presence and quality assessment, carbon storage and geothermal site screening, and rapid subsurface characterisation across frontier and mature basins.
Inputs: The model ingests post-stack seismic volumes and a broad set of derived seismic attributes spanning amplitude, frequency, phase, and structural domains. Typical inputs include RMS amplitude, envelope, spectral and instantaneous frequency measures, phase attributes, curvature, dip magnitude and azimuth, and other routine attributes commonly available in industry datasets. Client seismic data, attribute sets, and optional contextual inputs (e.g. basin setting, regional interpretation constraints) can be incorporated without requiring well data or prior labels.
AI Modelling Approach: The Exploration Model uses an integrated unsupervised and supervised learning framework. High-dimensional attribute space is first reduced using dimensionality reduction techniques to preserve geological signal while removing redundancy. Unsupervised clustering is then applied to identify statistically distinct and spatially coherent geological facies without introducing interpreter bias. These clusters can subsequently be used as training labels for supervised models, enabling prediction and transfer of facies patterns into new, unlabelled areas. The approach is designed to be scalable, reproducible, and adaptable across different geological settings, with minimal manual intervention.Outputs, Validation & Performance: Clients receive 3D facies and pattern volumes, cluster statistics, and spatially coherent geological classifications that can be directly loaded into standard interpretation software. Outputs highlight dominant depositional units, transitional zones, structural features, and high-amplitude or anomalous bodies relevant to exploration and reservoir assessment. Model performance is validated through spatial generalisation tests, cross-area deployment, and comparison against independent interpretations, demonstrating strong consistency and transferability with only limited manual refinement required. Delivery can be provided as raw seismic-compatible outputs or embedded within a GeonatIQ AI agent to support interactive analysis, rapid scenario testing, and iterative exploration workflows.
GeonatIQ’s Geochemical Exploration Model applies AI-driven geochemical analysis to accelerate ore body identification and target generation across igneous provinces and ore systems. The model is designed to convert legacy soil, regolith, rock-chip and drillhole geochemistry into ranked, defensible targets for follow-up mapping and drilling. Use cases include provincial screening, licence prioritisation, drill targeting, and rapid re-evaluation of historical datasets to reduce discovery cost and improve hit rates.
Inputs: The model ingests multi-element geochemical assays, element ratios and derived indices, lithology and alteration coding, structural interpretation layers, and spatial context. Where available, hyperspectral alteration products, remote sensing data, and geophysical layers (magnetics, gravity, radiometrics, EM/IP) can be integrated alongside client-provided reference targets such as known deposits, prospects, and barren domains. Public and network datasets can be layered in to strengthen regional context.
AI Modelling Approach: The model uses geology-respecting machine learning to identify coherent ore-system footprints rather than isolated assay highs. It combines supervised prediction, anomaly detection, and clustering to capture multivariate geochemical structure linked to igneous processes, alteration halos, and mineralisation styles. Spatially blocked train-test validation is embedded to control for autocorrelation and ensure that reported performance translates to realistic behaviour in unexplored ground. In most tabular geochemistry settings, our AI models provide robust, interpretable performance on highly imbalanced exploration datasets.
Outputs, Validation & Performance: Clients are offered prospectivity surfaces, anomaly domains, cluster maps, and ranked target inventories suitable for GIS workflows and technical review. Outputs are probabilistic, enabling risk-adjusted target comparison and scenario testing. Models can be retrained as new data arrives using a consistent core architecture developed by GeonatIQ in collaboration with Imperial College London.
GeonatIQ’s AI Lithium Price Model delivers short- to medium-term lithium price forecasts to support procurement strategy, contract negotiation, investment analysis, and risk management across battery materials markets. Use cases include supply-demand imbalance detection, price cycle analysis, project economics stress testing, and planning for miners, processors, battery manufacturers, OEMs, and investors.
Inputs: Our model ingests high-level signals including historical lithium prices, production capacity and ramp-up timelines, project approvals, inventories, battery demand indicators, EV adoption data, macroeconomic indicators, policy and subsidy signals, geopolitical and news data, and market positioning. Client-specific offtake data and proprietary supply chain intelligence can be incorporated.
AI Modelling Approach: Our model applies a meta-learning framework designed to adapt to structurally evolving markets characterised by rapid capacity expansion and demand growth. It integrates deep learning, signal decomposition, and adaptive training to capture price cycles, supply response lags, technological shifts, and regime changes across lithium chemicals and regions. A robust base architecture is maintained, with clients selecting additional lithium products, regions, or contract types to be trained on the same foundation.
Outputs, Validation & Performance: Clients receive multi-horizon price forecasts with confidence bands, directional and momentum signals, and scenario sensitivities reflecting demand growth and supply expansion risks. Outputs can be delivered as raw forecasts or via an AI agent for real-time market tracking and decision support. Forecasts are generated from a consistent core architecture retrained on a rolling basis and developed with Imperial College London. Client-selected signals and multiple lithium instruments (spot prices, regional benchmarks, and contract structures) are layered onto the architecture without rebuilding the model.
GeonatIQ’s AI Copper Price Model delivers short- to medium-term copper price forecasts to support trading, hedging, inventory management, and strategic investment decisions across physical and financial metals markets. Use cases include demand forecasting linked to electrification, supply disruption analysis, spread trading, and planning for miners, refiners, manufacturers, traders, and investors.
Inputs: Our model ingests high-level signals including historical copper prices, mine production and disruption data, refined output, inventories and exchange stocks, demand indicators linked to construction, electrification and manufacturing, macroeconomic indicators, geopolitical and news signals, freight and logistics data, and market positioning. Client-provided operational and commercial datasets can be integrated.
AI Modelling Approach: Our model applies a meta-learning framework that adapts dynamically to evolving supply-demand balances and industrial demand cycles. It combines deep learning, signal decomposition, and adaptive training to capture long-cycle behaviour, structural deficits, inventory drawdowns, and regime shifts in metals pricing. A robust base architecture is maintained, with clients selecting additional grades, exchanges, or regional benchmarks to be trained on the shared foundation.
Outputs, Validation & Performance: Clients receive forecasts across multiple time horizons, including confidence bands, directional and momentum signals, and scenario sensitivities tied to supply disruptions and demand shocks. Outputs are delivered as raw forecasts or via an AI agent for ongoing market monitoring and decision support. The base architecture is retrained on a rolling basis and developed with Imperial College London, with client-selected signals and multiple copper instruments (spot, futures, spreads, and regional benchmarks) layered on without rebuilding the model.
GeonatIQ’s AI EU ETS Carbon Price Model delivers short- to medium-term EUA price forecasts to support trading, compliance planning, hedging, and regulatory risk management. Use cases include allowance procurement strategy, compliance cost forecasting, policy stress testing, and trading for utilities, industrial emitters, financial institutions, and carbon funds.
Inputs: Our model ingests high-level signals including historical EUA prices, verified emissions data, compliance cycles, auction schedules, allowance supply dynamics, fuel switching economics, power generation mix, macroeconomic indicators, policy and regulatory signals, weather data, and market positioning. Client-specific compliance data and proprietary regulatory insights can be incorporated.
AI Modelling Approach: Our model applies a meta-learning framework designed to adapt to evolving regulatory structures and market behaviour. It integrates deep learning with signal decomposition and adaptive training to capture compliance-driven seasonality, policy-induced regime shifts, and structural changes in the carbon market. A stable base architecture is maintained, with clients selecting additional regulatory scenarios, fuel linkages, or contract structures to be trained on the same foundation.Outputs, Validation & Performance: Clients receive multi-horizon price forecasts with confidence bands, directional and momentum signals, and scenario sensitivities reflecting regulatory and policy change. Outputs can be delivered as raw forecasts or via an AI agent for real-time compliance monitoring and decision support. The model is retrained on a rolling basis using a consistent core architecture developed with Imperial College London. Client-selected signals and carbon instruments (spot EUAs, futures, calendar spreads, and policy scenarios) are layered onto the architecture without rebuilding the model.
GeonatIQ’s AI Gas Price Model delivers short- to medium-term natural gas price forecasts to support trading, hedging, portfolio risk management, and procurement planning across global and regional gas markets. Use cases include hub and basis trading, storage optimisation, LNG arbitrage, stress testing, and strategic planning for utilities, producers, traders, LNG operators, and industrial consumers.
Inputs: Our model ingests high-level signals including historical gas prices, regional supply-demand balances, storage and inventory levels, LNG flows, pipeline constraints, macroeconomic indicators, geopolitical and news signals, weather and climate data, and market positioning. Client-provided datasets and proprietary operational or commercial signals can be integrated.
AI Modelling Approach: Our model applies a meta-learning framework that dynamically adapts to changing gas market conditions, including seasonal effects and regional dislocations. It combines deep learning, signal decomposition, and adaptive training to capture temporal patterns, regime shifts, weather sensitivity, and structural changes across gas hubs. A robust base architecture is maintained, with clients selecting additional hubs, contracts, or spreads to be trained on the shared foundation.
Outputs, Validation & Performance: Clients receive forecasts across multiple time horizons, including confidence bands, directional and momentum signals, and scenario sensitivities, delivered either as raw outputs or via an AI agent for real-time monitoring and decision support. Forecasts are generated from a consistent core architecture retrained on a rolling basis and developed in collaboration with Imperial College London. Client-selected signals and multiple gas contracts (spot, futures, spreads, LNG indices, and regional hubs) are layered onto the architecture, enabling scalable coverage without rebuilding the model.
GeonatIQ’s AI Oil Price Model delivers short- to medium-term oil price forecasts to support trading, hedging, portfolio risk management, and strategic planning across physical and financial oil markets. Use cases include signal generation, hedge optimisation, stress testing, market outlooks, and planning for producers, consumers, traders, and investors.
Inputs: Our model ingests high-level signals including historical prices, supply-demandfundamentals, inventories and storage, macroeconomic indicators, geopolitical and news signals, weather and geospatial data, and market positioning. Client-provided datasets and proprietary signals can be integrated.
AI Modelling Approach: Our model applies a meta-learning framework that adapts dynamically to changing oil market conditions. It combines deep learning, signal decomposition, and adaptive training to capture temporal patterns, regime shifts, and structural changes. GeonatIQ maintains a robust base architecture, with clients selecting additional oil contracts or benchmarks to be trained on this shared foundation.
Outputs, Validation & Performance: Clients receive forecasts across multiple time horizons, including confidence bands, directional and momentum signals, and scenario sensitivities, delivered either as raw outputs or through a GeonatIQ AI agent for real-time monitoring and decision support. Forecasts are generated from a consistent core architecture that is retrained on a rolling basis and developed by GeonatIQ in collaboration with Imperial College London. Client-selected signals and multiple oil contracts (spot, futures, spreads, and regional benchmarks) are layered onto this architecture, enabling scalable coverage across markets without rebuilding the model.
GeonatIQ’s Commodity Price News Model is built around a large, structured news intelligence layer that quantifies how narratives, policy, and macro events influence commodity prices. It is explicitly context trained, not sentiment driven. The model is designed to understand how news affects prices depending on market structure and positioning, not how that news might be judged in a general or human sense. Context is everything. For example, war breaking out in the Middle East is clearly negative from a humanitarian or global stability perspective, but for an oil trader holding a long oil position the same event can be positive due to supply risk and price upside. Traditional sentiment models fail here. This model is built to handle that distinction. The approach was originally developed for carbon markets and is designed to generalise across commodities by retraining on commodity-specific historical data and outcomes. The framework was developed in collaboration with Imperial College London and has been submitted for academic publication, grounding it in rigorous research as well as real-world market application.
Inputs: The model is anchored on a curated database of approximately 280,000 news articles from top-tier sources including Bloomberg, Reuters, the Financial Times, AP, and similar outlets. For each commodity, historical prices are aligned with carefully labelled historical news, categorised by event type (policy, geopolitics, macro, supply, energy), expected price impact, volatility relevance, and time horizon.
AI Modelling Approach: The news layer is the foundation of the system. For each commodity, the model is retrained on historical price behaviour and aligned news labels to learn how different types of events have actually moved that market. Incoming news is then categorised against these learned patterns and fed into GeonatIQ’s numerical AI price prediction models. The numerical model generates the baseline forecast, while the news layer modulates that forecast based on current narratives, regime shifts, and event intensity.
Outputs & Use Cases: Clients receive multi-horizon price forecasts with directional signals, confidence bands, and attribution to specific news drivers. The model supports trading, hedging, risk management, stress testing, scenario analysis, and strategic decision-making, providing decision-grade context rather than simplistic sentiment signals.
