Executive Summary
A Fortune 500 integrated crude petroleum extraction operator, running mature multi-well pad assets across North America with tens of thousands of active wellbores and annual revenues exceeding 23B, struggled with fragmented production control and deteriorating well performance on aging Permian Basin assets. The company’s manual choke management workflow meant engineers were reacting to production losses rather than preventing them, costing millions in deferred and unrecovered EUR through infrequent adjustments. After deploying an AI-driven, closed-loop production platform capable of autonomous choke tuning at sub-minute frequencies, the operator recovered 14,400 incremental barrels per year and achieved 100,000-plus in annual efficiency savings, without adding headcount or new hardware.
Increase choke review frequency
Reduction in data reconciliation
Reduction in sensor data quality exception rate
SITUATION
The client is a vertically integrated North American crude petroleum producer with operations concentrated across the Permian Basin Delaware and Midland sub-basins and additional pad-drilling programs in the Rocky Mountain region. The company operates thousands of wells under multi-well pad configurations, where 8 to 16 horizontal laterals share surface infrastructure, separators, gathering headers, and compression, creating significant co-mingling and pressure interference between wells on the same pad. At the time of engagement, the company’s production optimization process was managed through a legacy SCADA system and periodic field engineer reviews. Engineers would physically assess or remotely monitor choke settings on a daily or weekly basis, leaving significant production optimization windows unexploited. With the average Permian well declining at 70-80% in the first year, the margin for optimization error compounds rapidly on mature assets.
TRIGGER
Three converging pressures forced the operator to act. First, a 2023 internal audit revealed that infrequent choke adjustments on maturing pad wells were contributing to measurable EUR underperformance relative to type curves; the gap between modeled and actual production widened as wells aged past 18 months. Second, commodity price volatility in late 2023-2024 tightened the operating margin, making every deferred barrel financially material. Third, competitors in the Delaware Basin, including peers adopting closed-loop AI platforms, began publishing efficiency gains that made the operator’s manual process a competitive liability. The status quo of once-daily or less frequent choke intervention was no longer defensible.
BARRIER
The operator faced a set of interconnected technical and organizational obstacles before deploying an automated solution.
Multi-well pressure interference: On 12-well pads, changing the choke on one well directly affects bottomhole flowing pressure on adjacent laterals. Manual engineers lacked the computational bandwidth to model real-time interference across the full pad.
Data silos and historian fragmentation: Production telemetry resided in multiple disconnected historians and SCADA systems across geographic regions. Engineers spent an estimated 1,700 man-hours per year manually extracting, normalizing, and reconciling data before any IT support or optimization action could be taken.
Infrequent intervention cadence: Manual choke reviews happened at best once per shift, leaving wells operating sub-optimally for hours at a time, a critical gap given that reservoir conditions and surface pressures fluctuate continuously.
EUR erosion from premature drawdown: Without real-time choke intelligence, some wells were drawn down aggressively in early production, causing near-wellbore damage and suppressing long-term EUR. Others were choked too conservatively, leaving immediate cash flow on the table.
No systematic optimization tracking: The company had no standardized workflow to evaluate, assign, or track the impact of field activities and choke workovers, making it impossible to quantify optimization value or prioritize engineer effort.
Implementation hurdle during discovery:
Approximately 22% of wellhead sensor data had systematic calibration offsets introduced by hardware replacements during COVID-era supply chain disruptions.
Solution
The IQ Platform, architected and deployed by UTOFA,was rolled out across the operator’s Permian Basin pad inventory in four phases: data integration, closed-loop control engine deployment, exception-based surveillance, and continuous model learning. It unified telemetry from three SCADA historians and the existing PI System, generated physics-based digital twins for all active wells, and delivered autonomous choke recommendations every 60 seconds, written back to field actuators within fully auditable control trails. IQ’s automated data quality engine detected and corrected COVID-era sensor calibration offsets before model training, preventing skewed baselines from compromising the closed-loop control system.
Why IQ Platform?
The operator chose the IQ Platform over RPA-based SCADA and DCA tools for its edge-native execution. Unlike cloud-dependent competitors that introduced latency, IQ’s microservices process telemetry and execute choke recommendations directly at the field level, no cloud round-trip, no optimization gap.
Technical Implementation
Data Integration Normalization (Weeks 1-4): IQ Platform connectors were deployed to unify telemetry from three separate SCADA historians and the operator’s existing PI System. Physics-based digital twins were generated for all active wells, incorporating wellbore geometry, reservoir pressure data, and historical production curves.
Closed-Loop Control Engine Deployment (Weeks 5-8): The IQ Platform’s autonomous choke optimizer was configured for each pad, establishing well-specific operating envelopes that balance immediate cash flow targets with long-term EUR preservation. The system analyzes live pressure, temperature, and flow telemetry alongside dynamic reservoir models to identify the production sweet spot, adjusting choke recommendations as frequently as every 60 seconds and writing commands back to field actuators within SOX-compliant audit trails.
Exception-Based Surveillance Layer (Weeks 9-10): Rather than replacing engineer judgment, the platform replaced screen-watching. Engineers receive alerts only when wells deviate from their modeled operating envelopes, freeing them to focus on higher-value intervention decisions rather than routine monitoring.
Validation Continuous Learning (Ongoing): IQ’s self-learning models ingest new production data continuously, recalibrating reservoir models as wells mature and pad interference patterns evolve. Production gains are tracked against a verified pre-deployment baseline, with performance reports structured to meet the operator’s strict internal audit and compliance requirements.
The Results
14,400 Incremental Barrels Recovered: 14,400 incremental barrels recovered per year, directly attributable to closed-loop choke optimization on mature pad assets, protecting EUR that would otherwise have been permanently deferred.
Choke Adjustments Every 60 Seconds: Choke adjustment frequency increased from once per-shift to every 60 seconds, eliminating multi-hour optimization gaps and keeping wells continuously at their modeled production sweet spot.
Data Quality Exceptions Down 86%: Data quality exception rate reduced from 22% to under 3%, automated sensor validation eliminated the manual reconciliation burden that previously consumed over 1,700 engineer-hours annually.
10x ROI in Year One: 10x ROI documented within the first operating year, consistent with published performance benchmarks for the underlying AI production platform across comparable North American asset classes.
Zero Headcount Addition: Zero incremental field headcount required, production engineers reallocated from routine screen monitoring to strategic well intervention, enabling each engineer to effectively manage a 40% larger well population.
Value in 10 Weeks: Deployed and produced value within 10 weeks, from initial data integration to live closed-loop control, with no CAPEX hardware additions or SCADA system replacement.
"The IQ Platform gave our production engineers the ability to manage twice the well count with the same team, not because we cut corners, but because the system handles the continuous optimization work and only escalates decisions requiring human judgment. That reallocation of engineering time is where the compounding value comes from."
"What surprised us was how quickly the data quality issues surfaced and were corrected. We had sensor drift we didn't know existed, the automated validation layer found it in the first week. Every model we had built previously was running on compromised baselines. Fixing that alone improved our type curve forecasting accuracy substantially."
Summary
- Achieved 14,400 incremental barrels recovered annually through closed-loop choke optimization on mature Permian pads.
- Increased choke adjustment frequency 99.8%, from once per shift to every 60 seconds.
- Reduced data quality exceptions 86%, from 22% to under 3% via automated sensor validation.
- Delivered 900% ROI in first year with zero headcount addition or CAPEX hardware.
- Cut data reconciliation time, eliminating 1,700 annual engineer hours previously spent on manual historian integration.
- Enabled 40% larger well portfolio management per engineer through exception-based surveillance.
Key Metrics
| KPI | Before IQ Platform | After IQ Platform |
|---|---|---|
| Choke review frequency | Once per shift (8–24 hrs) | Every 60 seconds (automated) |
| Annual incremental production | Baseline | +14,400 BOE/year |
| Data reconciliation (engineer-hours/yr) | 1,700+ hours | <100 hours (automated) |
| Sensor data quality exception rate | ~22% | <3% |
| Optimization tracking workflow | None (ad hoc) | Fully digitized, auditable |
| Deployment timeline | — | 10 weeks to live closed-loop control |
| ROI | — | 10x+ in Year 1 |
