Loading...
Loading...
The Bakken formation and its Three Forks bench made North Dakota the second-largest oil-producing state in the country — a position that held from 2012 through the early 2020s before slowly ceding ground to New Mexico's Delaware Basin surge. The Williston Basin that contains the Bakken extends across western North Dakota counties including McKenzie, Mountrail, Williams, and Dunn, with Tioga serving as one of the historic operational hubs of North Dakota oil infrastructure — Tioga is home to the Tioga Gas Plant, a major processing facility operated by Hess Corporation that processes gas and NGLs from thousands of Bakken wells across the basin. Hess, which has concentrated its Bakken position aggressively since divesting downstream assets, operates one of the most data-mature AI programs among mid-major Bakken producers. Continental Resources, founded by Harold Hamm and headquartered in Oklahoma City but with its operational heart in the Williston Basin, pioneered many of the horizontal drilling and completion strategies that made the Bakken commercial — and has been a significant AI adopter for completion optimization and production surveillance. Whiting Petroleum (now Chord Energy following its 2022 merger with Oasis Petroleum) rounds out the major independent producers in the play. The North Dakota Industrial Commission's Oil and Gas Division, headquartered in Bismarck, is the state regulatory authority overseeing drilling permits, production reporting, flaring compliance, and environmental monitoring — its public data systems are among the most detailed in any U.S. oil-producing state, providing rich training data for AI reservoir and compliance models.
Updated June 2026
The Bakken-Three Forks stacked pay in the Williston Basin creates a specific reservoir AI challenge that differs from Permian Basin multi-zone plays. The Bakken Middle Member and Three Forks are laterally continuous across millions of acres, but completion performance varies significantly with TOC content, natural fracture density, and distance from the Nesson Anticline. ML reservoir models trained on North Dakota Industrial Commission production data — which includes API well records, production by formation, and completion reports for 20,000+ horizontal wells — can predict EUR variance across proposed drilling locations with meaningful precision improvement over deterministic geological models. Hess has been particularly public about its AI-assisted completion optimization program, reporting that ML-driven perforation cluster spacing and frac fluid/proppant optimization across its core McKenzie and Mountrail County acreage has lifted average well productivity by 15-20% compared to pre-model completion designs. Continental Resources' completions team, working from its Williston operations base, uses production machine learning to rank in-fill drilling candidates across its ~1.1 million net Bakken acres in North Dakota — a portfolio too large for manual engineering prioritization. For smaller operators with Bakken positions ranging from 10,000 to 100,000 net acres, access to pre-trained Williston Basin ML models (using NDIC public data as the base training set) dramatically reduces the time-to-value compared to building proprietary models from scratch. The University of North Dakota Energy & Environmental Research Center in Grand Forks has published Bakken formation research that informs several commercial AI model validation efforts.
The Tioga Gas Plant — one of North Dakota's largest gas processing facilities, with capacity exceeding 300 MMcfd — represents the operational center of gravity for Hess's Bakken midstream integration. AI-driven SCADA monitoring at a facility of this scale addresses compressor health monitoring, separator efficiency optimization, and NGL fractionation yield maximization simultaneously. Computer vision applications at processing plants — automated inspection of rotating equipment, valve position verification from camera feeds, and early leak detection using thermal imaging integrated with SCADA dashboards — have reduced unplanned downtime at comparable facilities by 25-40% in documented case studies. North Dakota's Bakken midstream network, including ONEOK's Williston Basin gathering systems and the Hiland Crude pipeline network (now part of the Crestwood/Chord midstream complex), extends across hundreds of miles of rural terrain where manual inspection is expensive and anomaly detection latency can mean significant flaring events. The NDIC's flaring compliance program — which has progressively tightened gas capture requirements under the state's gas capture rules — creates direct regulatory consequence for midstream AI underperformance. Operators report that AI-assisted flare monitoring integrated with SCADA has been one of the highest-ROI NDIC compliance investments since the state's gas capture percentage rules went to 91% requirement in 2020. Integration competency with SCADA platforms used across North Dakota infrastructure — including Ignition by Inductive Automation, which has strong regional adoption among Bakken midstream operators — is a differentiating requirement for AI vendors in this market.
Exploration AI in the Williston Basin has evolved from early-stage Bakken de-risking — which is largely complete across the core area — to in-fill optimization and emerging upside identification in formations below the Three Forks, including the Sanish and Pronghorn benches that some operators are testing. Seismic interpretation ML, integrated with the NDIC's substantial public well and formation data, helps operators identify acreage where lateral lengths can be extended without disproportionate completion cost increase and where parent-child interference modeling suggests lower-risk in-fill spacing. Chord Energy, the merged Whiting-Oasis entity with one of the largest Bakken acreage positions, has emphasized AI production optimization as part of its post-merger operational integration — standardizing completion designs and production surveillance tools across what were two separate technology stacks. For the production optimization layer, AI surveillance tools monitoring the 17,000+ active Bakken horizontal wells in North Dakota face a specific challenge: many Williston Basin wells are beam-pumped after natural flow declines, and AI models that switch seamlessly between analyzing ESP performance data and beam-pump dynagraph analysis handle this asset lifecycle better than tools built exclusively for high-rate multi-zone plays. The Bismarck-based Petroleum Council of the NDIC serves as a regulatory liaison forum where AI-related production reporting and compliance topics increasingly appear on the agenda — operators who participate in this forum get earlier visibility into regulatory changes that affect AI compliance tool requirements.
Connecting AI systems to existing business infrastructure and workflows
Predictive models, data analysis, and ML pipeline development
Image recognition, object detection, video analysis, and visual inspection systems
Bespoke AI solutions, model fine-tuning, and custom model development