Digital transformation specialist Manish Kumar discusses how oil and gas manufacturing organizations are using structured operational data, intelligent automation, and enterprise analytics to improve operational visibility, workflow coordination, and large-scale decision-making.
A Shale Exclusive By Ellen F. Warren
Manish Kumar is Digital Project Manager at SLB in Houston, Texas, where he leads enterprise digital transformation initiatives focused on intelligent automation, operational analytics, AI-driven decision support, and digital governance across large-scale manufacturing operations. Over the course of a 15-year career spanning the United States and India, he has designed and deployed enterprise automation programs, analytics platforms, workflow-governance systems, and AI-enabled operational tools supporting manufacturing, supply chain, finance, quality management, and ERP environments.
His work has included building enterprise-scale automation programs that delivered measurable multimillion-dollar operational savings, developing predictive analytics systems for manufacturing coordination workflows, and leading digital transformation initiatives focused on operational traceability, data readiness, and AI integration across complex energy operations. In addition to his industry work, Kumar serves as Vice President of Technology for the Project Management Institute (PMI) Houston Chapter and frequently speaks on AI, automation, and digital transformation in energy and manufacturing operations.
In this interview, Kumar discusses the operational realities of deploying AI and intelligent automation systems inside oil and gas manufacturing organizations, including data governance challenges, enterprise workflow standardization, operational visibility, predictive analytics, and the evolving relationship between manufacturing operations and AI-enabled decision-making.
EW: You began your career working in business systems analysis, workflow automation, and enterprise process optimization before moving into large-scale digital transformation programs within the oil and gas (O&G) sector. What originally drew you toward operational systems and data analytics work, and how did you become interested in applying those capabilities inside complex energy and manufacturing environments?
MK: Early in my career, I kept running into the same situation: people knew a process was causing problems, but nobody could clearly see where the problem was occurring or how much impact it was having. Decisions were often based on assumptions because the underlying data either wasn’t available or wasn’t being used effectively. Working in business process management and workflow automation taught me that automation is really about understanding a process before you improve it. To automate something successfully, you have to understand every step, every handoff, every exception, and every dependency. That work appealed to me because it turns operational challenges into problems you can actually measure and solve.
My interest in energy and manufacturing came from the scale and complexity of the environments. These organizations generate tremendous amounts of operational data, but much of it is scattered across systems, spreadsheets, emails, and informal processes. A delay in a manufacturing operation is rarely an isolated event. It can affect production schedules, inventory availability, customer commitments, and downstream activities across the organization. I became interested in how better visibility, stronger data foundations, and more structured workflows could help organizations make better decisions and avoid those cascading impacts. Much of my work since then has focused on closing the gap between the information organizations already have and the information they can actually use.
EW: AI initiatives inside O&G manufacturing environments often struggle when operational data remains inconsistent, fragmented, or poorly governed. How do data quality and workflow standardization influence whether analytics and AI systems can produce reliable operational value at scale?
MK: Organizations often spend a great deal of time evaluating AI capabilities, but the quality of the underlying data usually determines whether those initiatives succeed. In manufacturing environments, operational information frequently resides across multiple ERP systems, spreadsheets, local databases, and manually maintained processes. Different teams may follow different standards, use different naming conventions, or maintain information with varying levels of discipline. Under those conditions, even sophisticated analytics tools can produce unreliable results.
One lesson I’ve learned is that data readiness deserves the same level of attention as the technology itself. During a major supply chain transformation, our team spent nearly six months reviewing, cleansing, and governing vendor master data before introducing automation. At the time, that work felt slow compared to the pressure to move the project forward. Once the SAP migration began, however, the benefits became obvious. Higher-quality data reduced downstream issues, improved consistency across processes, and created a stronger foundation for future automation and analytics initiatives.
EW: Your recent work has focused heavily on digital traceability and workflow visibility across large-scale manufacturing operations. What kinds of execution challenges and operational risks begin to emerge when coordination activities remain distributed across emails, spreadsheets, and disconnected communication channels?
MK: Distributed coordination creates a visibility problem that becomes more serious as operations grow. When requests, approvals, updates, and exceptions are managed through email chains and spreadsheets, it becomes difficult to understand where work stands, how quickly issues are being resolved, or whether operational commitments are being met consistently.
I encountered this firsthand in a warehouse operation that relied on more than 50,000 emails each year to coordinate activities. The organization maintained service-level commitments with logistics providers, but there was no reliable way to measure performance against those commitments. Delays often became visible only after they had already affected production schedules or customer expectations.
A second challenge involves organizational learning. When operational issues are handled one at a time through disconnected communication channels, recurring patterns can remain hidden for months. After we implemented a governed digital intake process, the operation gained visibility into request volumes, response times, compliance rates, and recurring issue categories. Information that had previously been scattered across thousands of individual email exchanges became available for analysis and continuous improvement.
EW: One of the recurring themes in industrial digital transformation is the challenge of standardizing workflows without disrupting ongoing operations. How do experienced teams introduce structured operational systems while still maintaining flexibility across fast-moving energy and manufacturing environments?
MK: The projects that tend to run into trouble are often the ones where system configuration begins before teams have agreed on how the process should actually work. Before technology becomes part of the discussion, stakeholders need agreement on workflow ownership, service expectations, escalation paths, and the way work will move through the organization. Those discussions may seem administrative, but they often determine whether a deployment succeeds.
The projects that have been most successful in my experience started by bringing operational stakeholders together to map existing processes and identify where different teams were approaching the same work in different ways. Establishing agreement on governance, responsibilities, and workflow design before configuring a system helps eliminate confusion later and creates much stronger adoption after deployment.
Energy, manufacturing, supply chain, and logistics operations require flexibility because teams are constantly responding to changing conditions. The challenge is creating enough structure to make work visible and measurable without making the process so rigid that people can’t adapt when circumstances change. A well-designed operational system should support both objectives.
EW: You have worked extensively across automation, ERP integration, analytics, and enterprise workflow design. How has the relationship between intelligent automation and operational decision-making evolved over the course of your career?
MK: Early in my career, automation was usually evaluated in terms of efficiency. The discussion focused on how many hours could be saved, how many manual tasks could be eliminated, or how quickly a process could be completed. Those benefits are real, but I eventually came to see a different source of value.
During a large finance automation initiative, a senior leader asked whether we could identify the types of transactions that were repeatedly failing validation and what those failures might tell us about upstream data quality. At the time, we couldn’t answer the question. The automation was processing transactions successfully, but we weren’t paying attention to the information contained in the exceptions.
That experience changed the way I think about automation. Exception data often provides a clearer picture of operational weaknesses than successful transactions do. Recurring validation failures, approval delays, and process exceptions frequently point to issues involving data quality, workflow design, or business rules that need attention. I now view automation as both an execution tool and a source of operational insight. The efficiency gains remain important, but the information generated by automated processes often helps organizations understand where problems originate and where improvements will have the greatest impact.
EW: Large energy and manufacturing environments generate enormous amounts of operational information every day. When organizations begin building predictive analytics capabilities, how do teams identify which operational signals are meaningful enough to support reliable decision-making and long-term operational planning?
MK: The first question I usually ask isn’t what data is available. I start by trying to discover what decisions people are struggling to make. Operational teams often have areas where decisions are based largely on experience because there is limited visibility into what is driving a particular outcome. Those situations provide a good starting point. Rather than asking what data is available, I prefer to ask where managers feel they are making assumptions, where unexpected problems occur, or where outcomes remain difficult to predict. Once those questions are clear, it becomes much easier to determine which data should be collected, structured, and analyzed.
In one warehouse analytics initiative, that approach revealed patterns that had gone unnoticed despite years of available operational data. Requests submitted late on Fridays experienced significantly higher breach rates than similar requests submitted earlier in the week. Analysis also showed that a relatively small group of high-volume parts accounted for a disproportionate share of service-level failures. In this case, neither finding required new data. The information already existed within the operation. What changed was the way the data was organized and analyzed to answer specific operational questions.
EW: Your work has included both enterprise-scale automation programs and smaller operational workflow initiatives. Over time, what patterns have you observed in digital transformation efforts that achieve sustained adoption and long-term operational value?
MK: I’ve seen projects deliver exactly what they were designed to deliver and still struggle to gain long-term adoption. Some of the most sophisticated systems I have seen delivered impressive functionality but struggled to gain long-term adoption because the organization never fully integrated them into daily decision-making. In most cases, the issue wasn’t the technology, it was that governance, accountability, and user adoption received less attention than implementation.
I’ve found that the initiatives that create lasting value share a few common characteristics. Stakeholders agree on governance before the technology is deployed, the system is designed around specific operational decisions and workflows, and analytics and reporting are built into the solution from the outset. Long-term adoption also depends on leadership behavior. Managers who consistently use operational data to guide decisions encourage their teams to do the same, which helps ensure that dashboards, reports, and analytics become part of everyday operations rather than tools that are consulted only occasionally.
EW: AI systems are increasingly being integrated into operational prioritization, escalation management, and real-time decision support across O&G operations. As organizations rely more heavily on AI-assisted workflows, which operational and governance safeguards become most important for maintaining reliability, accountability, and decision quality?
MK: One principle I consider essential is making sure AI systems operate within clearly defined boundaries. In manufacturing and energy operations, a confident but incorrect answer can create far more risk than an acknowledgement that the available information is incomplete.
In the environments where AI has performed consistently well, the underlying data sources are governed and the escalation paths are clearly defined. When sufficient information is not available, the system should identify that limitation and route the issue to the appropriate person instead of generating an answer based on incomplete information.
Human oversight also remains important, particularly when decisions carry operational, financial, or safety implications. AI can process information, identify patterns, and surface relevant data much faster than any individual can. Context, judgment, and accountability still belong to people. Used appropriately, AI helps decision-makers reach conclusions more quickly and with better information, while accountability for the outcome remains firmly in human hands.
EW: You have led digital initiatives involving manufacturing, supply chain, procurement, quality management, and operational coordination functions simultaneously. What kinds of coordination and ownership issues typically arise when enterprise analytics and automation systems are deployed across multiple business and manufacturing teams at scale?
MK: Most large transformation projects spend a lot of time discussing technology, but the tougher problems usually appear when multiple departments have to work through the same process. Large-scale initiatives often bring together teams with different priorities, different processes, and different definitions of success. Manufacturing, supply chain, finance, and quality groups may all be working toward the same business objective while maintaining separate workflows and governance structures. Once analytics and automation begin exposing issues that cross functional boundaries, questions about ownership and accountability quickly emerge.
One pattern I have encountered repeatedly involves problems that sit between two groups. An automated workflow or analytics platform identifies an issue, but responsibility for resolving it is unclear because the underlying process spans multiple functions. While those situations can create friction, they often reveal governance gaps that already existed before the technology was introduced.
The technology can identify the issue, but it can’t decide who owns it. Teams need agreement on responsibilities, escalation paths, and decision-making before those situations occur. When that groundwork is in place, bringing additional functions into the program becomes much easier.
Ellen F. Warren writes about industry leaders and trends in various sectors, including energy, fintech, IT innovation, healthcare, business, logistics, supply chain, commercial real estate, and entrepreneurship. As a former Independent Director, she served for more than a decade on the Boards of multiple E&P companies in the oil and gas industry.
Keep In Touch with Shale Magazine
As the new era of energy unfolds, you can bet we’ll be the boots on the ground to keep you informed. Subscribe to Shale Magazine for sharp insight into the arenas that matter most to your life. And don’t forget to listen to our riveting podcast, The Energy Mixx Radio Show, where our very own Kym Bolado interviews the most extraordinary thought leaders, business innovators, and industry experts of our time.
