The ChatGPT of Gas Leaks

Science-backed methane leak prediction using data companies already have.

November 2025

The Climate Impact: Why Methane Matters in Oil & Gas Operations

86×

Methane's warming impact over 20 years. Our most powerful short-term climate lever. ¹
Methane's warming impact (GWP20, CO₂e)

~12 years

in the atmosphere ²
Fastest pathway to cutting CO₂e impact.

95%

of natural gas is methane (CH₄) ³

17%

of global anthropogenic methane emissions come from Oil & Gas ⁴

¹ IPCC AR6 | ² NOAA Earth System Research Lab | ³ EIA Natural Gas Composition | ⁴ IEA Methane Tracker 2023

The Cost Impact: Expensive Losses from Invisible Leaks

~£30M

Estimated annual loss per mid and upstream operator from methane leaks (production losses up to 6%)

~3000

Mid-size operators globally (£0.5-20B revenue) facing ESG pressure and regulatory requirements ⁵

                        Operators lose money and reputation while contributing negatively to climate change, and current solutions don't scale to catch every leak.
                    

⁵ IBISWorld, S&P Global, Urgewald GOGEL, IEA

Why Now: The Perfect Convergence

Energy Transition

Natural gas is the bridge fuel. Operators face rising pressure to cut methane quickly — every molecule saved reduces emissions and increases revenue.

Tech Maturity

Cloud infrastructure, cheap compute, and scalable AI tooling have reached full maturity. Data centers, GPUs, and edge integration now make predictive emissions AI fast, affordable, and reliable.

Energy
Transition

Industry
Consolidation

The timing is ideal to deploy predictive AI for emissions.

Today's Approach Doesn't Scale

Today's Approach

HARDWARE-BASED (REACTIVE)

Satellites — expensive, only super-emitters, weather-dependent

Drones — non-continuous, weather-dependent, requires certifications

Aircraft — expensive, not continuous, heavily regulated

Handheld sensors — manual, slow, weather-dependent

Fixed sensors — millions to scale and maintain

VÖRNTEC's Approach

SOFTWARE-BASED (PREDICTIVE)

SCADA - Use the operational data operators already have.

AI - Predict leaks before they happen with AI pattern recognition.

Intermittent events - Catch intermittent events that hardware misses.

Cost - 24/7 coverage at a fraction of the cost.

                Methane leaks typically last 3–5 hours ⁶. To catch every leak with current methods, you'd need continuous sensors everywhere — costing millions.
            

⁶ Wang et al., 2022; Stokes et al., 2022

The Team: Built for Hard Problems

We spent a year validating this problem. Now we're ready to solve it.

WHO WE ARE

Matías — the artist

Founded Practical — embedded with O&G operators in Middle East

Understands operator workflows and field realities, helped tech and energy companies scale across GCC, Europe, and LATAM.

"Big picture to smallest detail"

Marco — the sailor

Founded MST Technology (IoT) — led ops at Oxford Quantum Circuits (pre-seed → Series C)

Built and managed mission-critical data infrastructure across NY, London, and Tokyo.

"Adaptable when context changes"

Operational insight + Technical design + Resilience

What We Do: We Turn Operational Data Into Predictive Intelligence

Detect leak patterns using AI before they occur.

INGEST

Connects to SCADA and ERP systems. Continuous monitoring of pressure, temperature, flow data.

PREDICT

ML algorithms estimate leak probability. AI interprets operational context like a domain expert.

ACT

AI Agent collaborates with teams via Microsoft Teams. Automatically creates work orders in ERP (e.g., SAP).

                        Industry Proof: Chevron found that 79% of detected events could have their source and duration determined using existing operational data ⁸. The insight is already in the system.
                    

⁸ Ravikumar et al., Environmental Science & Technology, 2020

Proven Model Accuracy: 99.58% on Peer-Reviewed Sensor Dataset

PROOF OF CONCEPT

We validated our ML approach on the UCI Gas Sensor Array Dataset — a peer-reviewed benchmark of real sensor data.

99.58%

Accuracy

1,912

True Positives

8 False Negatives (0.42%)
8 False Positives (0.14%)

Dataset: 7,680 readings across 4 gas types with 36-month sensor drift simulation

Model Performance

Performance across gas types and sensor drift conditions

This proves our ML approach works. Next: adapt it to SCADA data for methane leak prediction.

Built for Operators: AI That Collaborates With Your Team Seamlessly in Real Time

The best algorithm is worthless if operators don't act on it. We integrate directly into daily workflows.

Microsoft Teams – Maintenance / LDAR

VÖRNTEC Agent

🚨 Predictive Leak Alert

I've detected an anomaly with a high probability of turning into a leak:

📍 Location: Well #47 – Valve A-23
⏱️ Estimated time: 72 hours
🎯 Confidence: 94%
📊 Estimated emissions: 18 kg CH₄/day
💰 Financial impact: USD 1,200/day

Would you like me to create a preventive work order?

Javier Ruiz

Who is available to attend this valve today?

VÖRNTEC Agent

Available team:

👤 North Maintenance
👨‍💼 Supervisor: M. Fernández
📅 Today 14:00–18:00
🚗 12 km (≈20 min)

Javier Ruiz

Perfect. Create the work order with high priority and notify the supervisor.

VÖRNTEC Agent

✅ Action completed successfully

📋 WO #2025-0847 created in SAP PM
⚠️ Priority: High
📅 Today 14:00
👨‍💼 Assigned to: M. Fernández
✉️ Notification sent

The system will continue monitoring this valve.

EXPERT COLLABORATION

Like having the best domain expert working alongside your team — providing insights, coordinating actions, and ensuring nothing falls through the cracks. This isn't just reporting; it's a new way of working together.

From Lab Validation to Operational Deployment

Our validated approach translates directly to operational methane detection

DEPLOYED

Algorithms Validated

Multi-sensor fusion: 7 sensors → prediction
Drift compensation: Model adapts over time
Temporal patterns: 56 time-based features
Real-time: <100ms latency

AI Agent

Learns patterns: from operational data
Detects anomaly: predicts leak probability
Gathers context: pulls relevant data
Alerts via Teams: notifies operators
Opens WO in SAP: creates work order
Tracks to close: monitors resolution

NEXT MILESTONE

Validation

Train on operator data: SCADA (P, T, F, composition)
Label operator leaks: historical records
Adapt 56 features: to operations
Validate window: predict X hrs ahead

Foundation proven. BGV program gets us on track for the next impact.

Business Model: Simple, Scalable SaaS

SaaS Platform

Cloud-based AI agent connects directly to your existing infrastructure
SCADA integration for real-time monitoring
ERP integration (SAP) for automated work orders
Microsoft Teams integration for seamless collaboration

Multi-Agent Ecosystem

Specialized agents for environmental and cost impact:

Methane agent: Emissions monitoring and prediction
Water agent: Water usage optimization and leak detection
Sand agent: Proppant management and supply optimization

£15K/month

per AI agent

✓ No hardware costs
✓ No installation fees
✓ No maintenance overhead
✓ One prevented leak pays for 6+ months

Our Vision: The ChatGPT of Emissions

An open ecosystem where universities, companies, governments, and institutions connect to our models to build their own emissions solutions.

By standardizing how operational data becomes emissions intelligence, we unlock breakthroughs for the entire energy sector.

3.6M Cars

Equivalent petrol cars taken off the road for a full year, per operator ⁷

17 Mt CO₂e/year

Annual avoided emissions per operator

From single-operator use cases to a global platform for emissions intelligence.

We'd love to explore with BGV how to accelerate this path together.

⁷ U.S. Environmental Protection Agency (EPA). Greenhouse Gas Emissions from a Typical Passenger Vehicle.

How BGV Can Help Us

The Program

BGV provides: Network access + communications training + pre-seed fundraising preparation

Goal 1: MVP deployed with first operator, establish data pipeline
Goal 2: Refine pitch, connect with climate investors in BGV network
Goal 3: Demo Day preparation, secure second pilot commitment

Why BGV Specifically

Mission-aligned climate and impact network
Credibility to unlock data partnerships with operators
Technical mentorship for deep tech execution
Founder-first philosophy that aligns with how we build

We see BGV as a long-term partner in turning operational data into climate impact at global scale.

Thank You

Let's make synergies to turn operational data into climate impact.

Contact

marco@vorntec.com