Scott Smidlein
Senior Client Executive
-
United States
This article is co-authored by Marsh and Solomon Associates.
Midstream and downstream energy operators face a pressing operational imperative: extracting greater yield from aging assets while preventing high‑severity incidents that can disrupt supply, harm operators, and damage reputation.
As they work to refine crude oil into fuels, chemicals, and other finished products and transport and store these products, many midstream and downstream operators are under increasing pressure to optimize performance and efficiency. At the same time, they may face a major challenge — evolving their operations within an infrastructure built for a different era.
Midstream and downstream operators’ important role in the oil and gas value chain means these entities often occupy a pivotal place in our nation’s economy. But their operations are often limited by older plants and pipeline networks that were designed for and operated through processes that relied on manual inspections, episodic data capture, and procedural training.
Today’s operating context is different: markets move faster, margins are tighter, the workforce is evolving, and tolerance for unplanned downtime is low. To improve operational efficiency as well as safety outcomes, midstream and downstream operators should consider investing in new technologies that can help improve decision-making, reduce operational risk, and enhance workforce capabilities.
Historically, midstream and downstream operators have relied on periodical reporting, rule-based alarms, and operator experience to guide production-related decisions. Financial insights were often reconciled once production was complete and plant-level analytics were coarse and often delayed. The result was a slow reaction to shifts and missed margin opportunities. Pipeline inspection was labor-intensive, mainly conducted through helicopter flyovers, pigging runs, and manual thermal checks, which were often slow to identify emerging integrity issues.
Today, operators have the opportunity to integrate real-time analytics with market data to make faster production decisions. Artificial intelligence (AI) models, for example, can rapidly recommend product slate shifts that optimize margin, avoiding hours or days of guesswork. Automated inspection technologies, such as drones, ground‑penetrating infrared surveys, and intelligent inline inspection tools, can supplement traditional pigging and aerial surveys to provide more frequent, higher-resolution condition data. AI‑driven pattern recognition software can highlight anomalies that might be missed through a human review and turn logs into actionable maintenance alerts. While these capabilities do not replace skilled operators, they can augment decision‑making and surface issues earlier so field teams can act.
Modernize data hygiene and connectivity, starting with an inventory of processes and inspection data sources, standardizing them where possible. Consider piloting AI use cases that are tied to clear commercial outcomes, such as margin improvement following a product switch, and measuring results.
Maintenance has long been reactive or calendar-based, revolving around scheduled turnarounds, routine visual inspections, and periodic testing. Capital planning was often forced to prioritize reactive fixes or blanket refurbishments rather than targeted investments that were most needed. Aside from the potential of unnecessary spending, this approach could leave operators exposed to unseen failures. Solomon’s Reliability and Maintenance (RAM) Study benchmarking data bears this out: top-quartile (Q1) performers typically spend less than 5% of maintenance work hours on emergency corrective work, compared to nearly 15% for Q4 performers, while dedicating significantly more time to preventive and predictive activities (Solomon Associates, 2026).
Figure 1 (Solomon Associates, 2026)
A shift to condition-based and predictive maintenance can materially reduce unplanned downtime and optimize capital deployment. When combined with trend analytics, high-resolution monitoring — using thermal imaging, vibration analysis, and corrosion sensors, among others — can provide early warnings of deterioration.
Predictive algorithms that account for seasonal variation and operational cycles can help identify the subset of assets that will benefit most from intervention, enabling prioritized spending and more efficient turnaround planning. For pipeline operators, remote infrared inspection and geospatial monitoring reduce the need for intrusive checks and enable faster, safer responses to emerging threats.
Start by mapping critical equipment and deploying sensors at high-value or high-risk nodes to demonstrate value. Integrate sensor outputs into maintenance management systems and use small, focused pilots to refine predictive models. Consider rolling predictive insights into capital expenditure planning so that investments are risk‑informed and aligned with lifecycle economics.
Health and safety programs have traditionally focused on classroom instruction, checklists and on‑the‑job mentoring. While essential, these methods can leave knowledge gaps, particularly for infrequent, high‑consequence tasks for which hands‑on exposure in a live environment can be unsafe or impractical. Similarly, legacy gas‑detection systems provided point readings but limited spatial coverage, often sending alerts when an incident was underway, and safety might have already been compromised. These issues, combined with an evolving workforce, present a real threat to safe and reliable operations.
Immersive technologies, such as virtual reality, can enable realistic, repeatable training in risk-free settings. Operators can rehearse infrequent but high‑risk procedures — for example, valve closures, emergency isolation, or manual shutdown steps. Repeated exposure to virtual training better enables crews to internalize the necessary steps before they encounter these situations in the field.
Meanwhile, modern sensor networks and distributed detection systems, including radar‑profiled gas detection and wide‑area sensor grids, offer earlier alerts and broader situational awareness, giving teams more time to respond and protecting nearby communities and the environment. When VR training is paired with documented procedures and scenario‑based drills, it demonstrably increases proficiency without putting personnel at risk.
Introduce virtual reality or simulator‑based modules for critical shutdown and emergency procedures, and evaluate their efficacy through measured competency assessments. Expand gas‑detection coverage at key interfaces and integrate sensor feeds with incident‑management systems that translate alerts into coordinated, practiced responses. Ensure training and technological upgrades are integrated into existing safety management systems — and documented for regulators and insurers.
Technological investments that enable preemptive risk mitigation strategies can be shared during insurance renewal meetings to improve an organization’s chances for better rates and terms and conditions, especially when there is a demonstrable reduction in the frequency or severity of losses.
However, human oversight remains key and is often one of the questions that insurers ask, together with inquiring how new tools are embedded within procedures.
Operators that can demonstrate data governance, strong verification protocols, and clear ways to integrate alerts into work processes can gain top-tier status that paves the way for conversations with insurers that can lead to improved terms.
Senior Client Executive
United States
HSB Solomon Associates LLC, www.SolomonInsight.com.
The opinions expressed herein are valid only for the purpose stated herein and as of the date hereof. Information furnished by others, as well as public information and industry and statistical data, upon which all or portions of this publication may be based, are believed to be reliable but have not been verified. We have used what we believe are reliable, up-to-date and comprehensive information and analysis, but all information is provided without warranty of any kind, express or implied, and we disclaim any responsibility for such information or analysis or to update the information or analysis in this publication.