During the pandemic, I became involved with a company that was rethinking the work of the quality professional at fabrication shops. The traditional practice was to dispatch a company person to the fabrication business to see what was being built to make sure it conformed to the specification. The pandemic rendered site visits hard to pull off, but the quality standards could not be relaxed.
Such equipment manufacturers have to comply some of the strictest standards of any sector. It’s still surprises me that many companies manage the delivery of quality through spreadsheets, paper documents, disconnected folders, and manual processes. Systems like these are what technical people call fragile. They don’t cope well with change, they’re a pain to maintain, expensive to run, and are dangerously dependent on the knowledge of a handful of insiders.

When a defective part turns up, particularly if it has caused an upset, it’s then a five alarm fire trying to find the cause. Inspection records, supplier documents, and production histories may be scattered across multiple systems. And heaven forbid that the quality manager retires, or leaves for better opportunities.

You might not be surprised that the fragile quality system environment is attracting the attention of AI as a solution. Modern quality management platforms can connect previously disconnected processes, automate documentation, track corrective actions, and make quality information available through natural language conversations. Rather than hunting for the elusive records related to an obscure part, quality managers can just ask their systems questions, issue instructions, and focus on improving operations.

In this episode, I’m speaking with Onega Ulanova, co-founder of QMS2GO, and a former American Petroleum Institute auditor. We discuss the shortcomings of traditional quality management, the potential of AI-powered quality systems, and how better compliance can improve reliability, productivity, and cash flow.

👤 About The Guest

Onega Ulanova is the co-founder of QMS2GO, a company developing AI-powered quality management solutions for industrial manufacturers and service providers.

Her career includes roles as a quality manager, consultant, and lead auditor for the American Petroleum Institute, where she audited more than 200 facilities worldwide. She has extensive experience with API Q1, API Q2, and ISO 9001 quality management standards.

Together with her husband and business partner, Konstantin Dolgan, she is applying this experience to make quality management more accessible, automated, and effective.

Artificial intelligence has been part of the oil and gas industry's technology toolkit for years. Predictive models identify deteriorating equipment, flag production anomalies, and help engineers make better decisions. These tools generate valuable insights across exploration, production, and maintenance operations.
But knowing something is wrong and doing something about it are two different things. Someone still has to interpret the information, determine the appropriate response, and organize the work. Frustratingly, highly skilled specialists spend enormous amounts of time gathering and cleaning data, the most experienced operators are rapidly approaching retirement, and industrial facilities remain difficult and expensive to inspect and maintain.
I’ve speculated in my books that AI could someday move beyond generating predictions to actually helping carry out the work. Finally, that day is here.
In this episode, I speak with Alex Knowles, Industry Chief Technology Officer at DataRobot, about the emergence of agentic AI in oil and gas. Drawing on his experience in the US Navy, oil and gas operations, and enterprise AI, Alex explains how intelligent agents can help bridge the gap between insight and action.

• compressing months of work into days

• preserving the knowledge and experience of senior people

• optimizing maintenance across complex industrial assets

• collaborating with Chevron to develop AI-enabled autonomous inspections

• using physics models and sensor data to predict hazardous gas plumes and guide robotic inspections

I'm most interested in where agentic AI is already delivering value, why annoying tasks such as cleaning well data and seismic metadata deserve more attention, and what happens when AI begins to interact with the physical world of oil and gas.

⚒️ Additional Tools & Resources

📚 My books: https://geoffreycann.com/book/

🎬 Go backstage and check out my studio:https://geoffreycann.com/mystudio/

🎓Take my one-day digital strategy training course for oil and gas: https://www.udemy.com/course/digital-oil-and-gas/?referralCode=0161D4D49AB75735A185

-----------------------------------------------------

🔗 Connect with Me

📘Blog series: https://digitaloilgas.substack.com/

🎧 Podcast: https://geoffreycann.com/podcast/

💼 LinkedIn: https://www.linkedin.com/in/advocate-digital-innovation-for-energy/

🐦 X: https://x.com/geoffreycann👉 https://geoffreycann.com/contact/

-----------------------------------------------------

⚠️ Disclaimer

The views expressed in this podcast are my own and do not constitute professional advice.

How Infrared Spectroscopy Is Changing Industrial Water Diagnostics, an interview with Graham Roeber of Orono Spectral SolutionsOil and water don't mix, but in oil and gas operations, they frequently end up together. Refineries, processing plants, and other industrial facilities routinely test their water streams for hydrocarbon contamination. Historically, this testing has been driven by environmental compliance. The objective is to demonstrate that discharged water meets regulatory standards, rather than to improve the performance of the facility.The problem is that traditional oil-in-water testing can be slow, labour-intensive, and expensive. Results can take hours, making the technology poorly suited to diagnosing leaking equipment or monitoring process performance. Meanwhile, oil escaping into wastewater represents lost product, increased treatment costs, and potentially costly equipment failures. For operators trying to improve efficiency without adding capacity, these losses matter.Infrared spectroscopy offers a different approach. By capturing oil from water samples on a specialized membrane, the technology can measure hydrocarbon concentrations in minutes rather than hours. More intriguingly, the spectral fingerprint of the captured oil can be compared against reference data to help identify the type of oil present. That makes it possible to narrow down the source of contamination, detect emerging problems, monitor process trends, and make better operational decisions.In this episode, I speak with Graham Roeber, CEO of Orono Spectral Solutions (OSS), about how this technology is transforming industrial water diagnostics. We discuss how chemical fingerprints can help maintenance teams locate leaking equipment, why faster testing changes the economics of process monitoring, the challenges of gaining regulatory acceptance, and how AI is making sophisticated chemical analysis more accessible. We also explore emerging applications in desalination, ammonia, corrosion monitoring, and other industrial processes.👤 About the GuestGraham Roeber is CEO of Orono Spectral Solutions (OSS), a company specializing in infrared spectroscopy technology for oil-in-water analysis and industrial water diagnostics.OSS develops analytical solutions that help industrial operators measure hydrocarbon contamination, identify oil types, and gain actionable insights into the performance of their equipment and processes.Website: https://www.ossmaine.com/Graham on LinkedIn: https://www.linkedin.com/in/grahamroeber/#DigitalOilAndGas #InfraredSpectroscopy #WaterTechnology #IndustrialInnovation #OilAndGas #OronoSpectralSolutions-----------------------------------------------------⚒️ Additional Tools & Resources📚 My books: https://geoffreycann.com/book/🎬 Go backstage and check out my studio:https://geoffreycann.com/mystudio/🎓Take my one-day digital strategy training course for oil and gas: https://www.udemy.com/course/digital-oil-and-gas/?referralCode=0161D4D49AB75735A185-----------------------------------------------------🔗 Connect with Me📘Blog series: https://digitaloilgas.substack.com/🎧 Podcast: https://geoffreycann.com/podcast/💼 LinkedIn: https://www.linkedin.com/in/advocate-digital-innovation-for-energy/🐦 X: https://x.com/geoffreycann👉 https://geoffreycann.com/contact/-----------------------------------------------------⚠️ DisclaimerThe views expressed in this podcast are my own and do not constitute professional advice.