In conversation with Keith Davies, Prometheus Group’s Chief Technology Officer

Q: To start, could you give us a brief overview of Prometheus Group and its main focus areas?

Keith: Prometheus Group is a global enterprise asset management solution provider helping asset-intensive industries operate more safely and efficiently. We really “grew up in the plant”.  Our story began with understanding the everyday challenges of maintenance and operations teams. From those early days, our focus has always been on improving user experience and helping people get their jobs done faster, safer and with higher quality.

Over time, we’ve expanded from routine maintenance scheduling into a much broader suite of solutions across the entire asset lifecycle. Today, we’re proud to serve sectors such as oil and gas, power and utilities, mining, chemicals, pulp and paper, and food and beverage.

Q: You often emphasise that everything starts with data. Why is that such a central part of your AI vision?

Keith: That’s right. It all starts with data. When we begin to talk about our AI platform vision, it doesn’t start with AI frameworks or agentic models, even though those are critical pieces of what we do. For us, it begins with data because that’s where we’ve always differentiated ourselves.

We’ve been working with customers for more than 20 years to cleanse and enrich their data. When we first started, we noticed that across industries – from oil and gas to utilities and manufacturing – companies struggled with incomplete or duplicate data that was costing them hundreds of millions of dollars annually.

So, we built solutions that would give them cleaner, more consistent, and more usable data. That foundation now underpins our AI platform. Because our platform’s footprint touches nearly all aspects of maintenance and operations data across plants, when we connect that with a customer’s data, we’re creating a truly unique and comprehensive data view for them.

AI models evolve rapidly, but one thing remains constant: the quality of results always depends on the quality of the data. Clean, complete data is what enables AI to deliver meaningful, actionable insights.

Q: How have you built your AI platform on top of that data foundation?

Keith: Building on a strong data foundation, we wanted our AI to be highly deterministic, meaning that it doesn’t “hallucinate” or make guesses when the data isn’t there to support an answer. We designed our platform with that in mind.

We’ve implemented guardrails and fine-tuning mechanisms, so the AI only responds when it has sufficient, verifiable data. On top of that, we’ve added robust security and user-permission controls to ensure the system only performs actions users are authorised to execute.

From a functional standpoint, we’ve developed two kinds of AI agents within our platform. Assistive agents respond to user input to complete specific tasks or workflows. Autonomous agents, meanwhile, proactively make recommendations based on operational data. Both types share one core philosophy: AI never acts without user confirmation.

This design keeps users firmly in control, combining automation and human oversight to deliver reliability and trustworthiness – qualities essential in industrial environments.

Q: Can you share some examples of how the AI platform is being used to solve real customer challenges?

Keith: Certainly. One of our key focus areas has been automated maintenance scheduling. Using AI, we can build resource-constrained schedules in seconds – taking into account business priorities, compliance needs, and production impacts. The platform can identify risks by flagging material readiness, permit status, or other dependencies, allowing maintenance teams to act before those issues disrupt operations.

We’re also applying AI in automated planning, where the system leverages historical asset data and performance norms to create planned corrective orders automatically. From there, AI can generate preventative and predictive maintenance recommendations, helping teams move from reactive to proactive asset management.

In short, the platform reduces manual work and complexity while increasing operational insight. It helps customers focus their limited resources on the right work at the right time to improve wrench time, safety, and asset reliability.

Q: You mentioned earlier that Prometheus has deep roots in the plants. How does that influence how your AI operates?

Keith: That’s a big part of what makes us different. Many companies try to apply generic AI tools to industrial data, but they often lack an understanding of the real-world operational constraints and processes inside plants.

We’ve spent decades working alongside maintenance teams, planners, and schedulers. We know their workflows, pain points, and what “good” looks like. That domain knowledge is what shapes our AI’s behaviour. It’s not just processing data; it’s understanding how that data relates to real industrial decisions.  In fact, we’ve fed these best practices into our AI models, so they’re incorporated into the decision logic.  It’s yet another thing that differentiates us.

That’s why our AI doesn’t try to reinvent what works, but rather it accelerates what users are already doing. It calls existing functions within our applications, helping users complete tasks faster, with fewer errors. It’s about amplifying capability, not replacing expertise.

Q: How does Prometheus approach AI deployment in different regions, particularly here in the Middle East?

Keith: The Middle East is an exciting region for us. Companies here recognise the value of data and are eager to use it more intelligently. One important distinction, however, is that many organisations prefer to keep their data within their national boundaries.

Some have built private clouds to ensure data sovereignty, while others partner with public-cloud providers. We designed our AI platform to be cloud-agnostic, so it can run seamlessly on AWS, Google Cloud, or Microsoft Azure – and it can also be deployed on a customer’s private cloud.

That flexibility is key. It allows companies in MENA to adopt AI according to their own IT strategies and data-governance policies.

Q: Could you share any examples or early feedback from customers in the region?

Keith: We have a number of customers in the Middle East and are actively engaging with major energy and industrial companies in the region. Just recently, we presented our AI platform to a consortium of three of the world’s largest oil and gas producers, and the response was very positive. They were able to use the platform directly and see its capabilities first-hand. The ADIPEC conference will be another major opportunity to collaborate with the MENA region’s leading operators on AI-driven asset optimisation.

We’re also seeing strong interest from utilities and other process industries. Many organisations in the Middle East already have predictive-maintenance software, but they see AI as the next logical step, taking them from insight to action and from analysis to automation.

Q: Finally, how would you describe the overall receptivity toward AI in this region?

Keith: Very strong. Companies like ADNOC are clear examples – they’ve established dedicated AI budgets and innovation programmes. That level of commitment signals that organisations here see AI not as an experiment but as a core enabler of their operational strategy.

We see this as a perfect opportunity to partner with these organisations to bring practical, results-driven AI solutions to life – solutions grounded in Prometheus Group’s long-standing commitment to data, safety, and efficiency.

Prometheus Group is currently at Adipec. To arrange a meeting, please click here.

Published in partnership with

Prometheus Group