Back to Oil & Gas Hub
Upstream

Production & reservoir surveillance

One of many upstream use cases. This example focuses on well optimization and decline curve analysis, turning daily production data into actionable surveillance intelligence.

The surveillance gap

Your production engineers monitor hundreds of wells across multiple pads. Daily volumes arrive in spreadsheets. Decline curves are modeled in siloed tools. When a well underperforms, the question isn't whether you have data. It's whether anyone noticed in time.

The gap between measurement and action costs operators millions in deferred production annually. Not because the signal isn't there, but because no one had the bandwidth to look.

3-10%
Throughput increase from AI-driven production optimization (McKinsey)
48hrs
Average delay before underperformance is flagged
100s
Wells per engineer: too many to watch manually

Why current tools fall short

Traditional approaches were not built for the scale and complexity of modern operations.

01

Spreadsheet-based decline curves

Arps models in Excel are static. They don't auto-update, don't flag deviations, and can't correlate across wells or reservoirs.

02

Delayed production accounting

Deferred production is often identified weeks after the event. By then, the well may have been offline for days without intervention.

03

Disconnected data sources

SCADA, production accounting, lab data, and well tests live in different systems. Correlating them requires a data engineer and a ticket.

04

One-off analyses

Each investigation is a custom project. The analysis isn't reusable, and the knowledge leaves when the engineer moves on.

The Lumina Approach

How Lumina solves it

Three layers of intelligence working together: reasoning agents, proactive detection, and multi-agent deliberation.

Layer 1

Decline curve intelligence

Lumina agents ingest daily production volumes, fit Arps models (exponential, hyperbolic, harmonic), and flag wells deviating from expected decline. Every calculation runs as SQL against your data. No black box.

Every calculation is SQL you can verify. No black box.

Agent Reasoning
SQL-Backed
Agent Reasoning:

Analyzing Well PAD-07-12... Fitting hyperbolic decline: Qi=850 bbl/d, Di=0.12, b=0.8 Current rate 620 bbl/d vs expected 710 bbl/d → 12.7% underperformance detected → Correlating with water cut trend: WC increased from 45% to 62% over 30 days

All outputs backed by verifiable SQL you can inspect
The Radar
Scanning
Anomaly Detected:

ALERT: Cluster of 4 wells on Pad 07 showing simultaneous GOR increase (+18% avg over 7 days) Hypothesis: Gas cap breakthrough in shared drainage area Confidence: HIGH (p=0.003 via correlation analysis) → Recommend: Reservoir team review pressure communication study

Proactive hypothesis testing, like anomaly clustering
Layer 2

Proactive well surveillance

The Radar continuously scans production data for anomalies: sudden rate changes, water cut excursions, GOR spikes, ESP current deviations. It surfaces issues before they become incidents.

The Radar surfaces issues the operator didn't know to look for. Before they become incidents.

Layer 3

Multi-agent production review

Production, reservoir, and facilities agents deliberate on underperforming wells. Each brings domain-specific reasoning: decline analysis, reservoir pressure, and surface constraints.

The output is grounded in facts (SQL results), not hallucination. Every recommendation carries a full audit trail.

The Boardroom
Deliberation
Multi-Agent Debate:

Production Agent: Well PAD-07-12 underperforming 12.7%, recommend workover Reservoir Agent: Offset well PAD-07-11 shows same trend, this is drainage, not mechanical Facilities Agent: Separator pressure increased 15 PSI last week, backpressure effect → Consensus: Adjust separator pressure first (low-cost). Monitor 7 days before workover decision.

Agents vote, challenge, and produce a synthesized recommendation

The result: intelligence that scales

Lumina addresses the four strategic problems that hold operators back.

Scale without headcount

Monitor hundreds of wells per engineer. Agents run surveillance 24/7, flagging only what matters.

Reduce reactive costs

Catch deferred production days earlier. Reduce unnecessary workovers by correlating root causes across domains.

Knowledge retention

Decline models, well behavior patterns, and operational playbooks are captured in the system, not in someone's head.

Auditable trust

Every rate calculation, decline fit, and anomaly score is backed by SQL you can verify.

Agents for this use case

Specialized AI agents that power this workflow.

Lumi Prod

Production Analyst

Specializes in decline curve analysis, rate forecasting, deferred production accounting, and well test validation.

Lumi Res

Reservoir Surveillance

Monitors water cut trends, GOR behavior, pressure communication, and drainage patterns across well groups.

Lumi Core

Data Scientist

Handles data ingestion, schema detection, unit conversion, and cross-system correlation.

This is just one of many use cases

Explore what Lumina can do for your operation

Production & reservoir surveillance is one example of how Lumina reasons on operational data. Across Oil & Gas, every domain has use cases where AI agents can add value.

Not ready to commit? Stay up to date as we release new capabilities and industry-specific agents.

View all Oil & Gas use cases