Navigating the IT/OT divide: Strategies for successful AI project implementation

I have previously written about the opportunities that lie in leveraging AI for the process industry. As the article covers, the process industry could reap great benefits from using AI to improve resource utilization, prevent unplanned downtime, increase capacity and quality, and minimize its environmental impact. However, there are some hurdles when it comes to succeeding with AI projects and initiatives in the industry. One of them is the “IT/OT divide”.

If you work in the industry, this is most likely a known topic. For me, it was an eye-opener when I moved into the industrial software space in 2018. I have an Information Technology (IT) degree, and during my 20+ years in the IT software industry I never really encountered the term Operational Technology (OT) until I joined an Industrial DataOps platform company. As I started to delve into the OT side and talk with industrial clients, it became clear that there is indeed a divide, and that it stalls the implementation and scaling of new technologies, and with it the business value. Before diving into how to bridge the gap, let’s first take a closer look at what IT/OT means and why convergence is necessary.

The IT/OT divide

The IT/OT divide refers to the separation between the IT and OT domains within an organization. The two domains have traditionally served distinct purposes, with different focuses, systems, and expertise.

Historically, these two domains operated relatively independently, each with its own technologies, standards and objectives. With the rise of Industry 4.0 and the increasing integration of digital technologies in industrial processes, there is a growing recognition that the gap needs to be bridged. Collaboration between IT and OT is crucial for organizations looking to implement advanced technologies like AI to improve efficiency, productivity and overall operational performance.

Technological advancements enhance IT/OT convergence possibilities

There is no playbook for successful interoperability and interaction between IT and OT. However, technological advancements are making IT/OT convergence more achievable than ever before. Convergence is not about turning IT professionals into plant engineers or machine operators into data scientists. It is about a strategic approach to align and unite subject matter experts, data and solutions that have previously been isolated in OT and IT teams.

The timing to act on IT/OT convergence has never been better. Mature industrial enterprises are:

• realizing maintenance and production optimization by putting Industrial AI in the hands of their process engineers
• getting out of “pilot purgatory” and scaling business value with Industrial Software-as-a-Service (SaaS)
• democratizing access to OT data through standardized and secure integrations

The role of industrial AI SaaS solutions

The benefit of Industrial AI software built for process engineers is that the people who know the process best can create and use machine learning models themselves, without programming expertise or a data science team. An Industrial AI SaaS solution offers a different way to overcome the challenges of traditional development methods. Here’s how:

Here's how:

1. Empowering Domain Experts: In my years working with industrial organizations (not saying it happens in your company), I’ve seen the IT/OT gap play out like this: the operational side feels that IT doesn’t understand the production processes well enough to know what they need, and IT feels that OT doesn’t appreciate the value of what they deliver. Industrial AI software built for industrial processes and industrial users can ease this to the benefit of both sides. Domain experts get intuitive tools to build, fine-tune and use machine learning models themselves, without programming knowledge or extensive resources from IT. IT, in turn, can provide a solution that is tailored to industrial needs and quickly lets the business realize value. Letting domain experts solve operational challenges in real time, every day, ensures that AI becomes a daily decision-support tool for engineers and operators.

2. Accelerating Time to Value: Most AI initiatives have historically been IT-driven do-it-yourself (DIY) projects, built from scratch and often requiring custom code. DIY projects can generate results, but implementation often stretches over years rather than months, and they can be hard to scale. How do we adapt the models for the next production line? And the next plant? Who takes responsibility for maintaining them? The significant cost of a DIY platform is often not building it, but maintaining it. A SaaS solution offers continuously updated machine learning capabilities and scalability without IT having to recode for each new scenario. It can be set up, configured and deliver operational value in less than a month, and scale easily across the organization. As a bonus, it removes the need for expensive hardware and the ongoing maintenance costs normally associated with industrial software.

3. Unlocking data: A modern Industrial SaaS solution allows rapid integration with OT systems through pre-built connectors for standard industrial protocols and systems, as well as existing cloud data platforms. Building and maintaining custom integrations takes significant time and resources from IT. With an Industrial SaaS solution, IT can focus on security and data integrity across the enterprise, while operations realize business value from easily accessible OT data.

The people side: from handover to shared ownership

Technology can lower the barriers, but in my experience the biggest obstacle to closing the IT/OT gap is not technical. It is ownership.

IT and OT are measured on different things. IT is responsible for security, standardization and keeping systems running across the enterprise. OT is responsible for safety, uptime and hitting production targets at the plant. Both are right to protect what they own. But when an AI project starts, this often means that each side sees the other as a risk rather than a partner. IT worries about new connections into the control network. OT worries about tools that don’t understand the realities of the process. And when the pilot is over, the question nobody has answered is: who owns this now?

The organizations I’ve seen succeed do three things differently:

1. They agree on the problem before the technology. Starting with a concrete operational challenge, like an unstable unit, a quality issue or a recurring stop, gives both sides a shared goal that is easy to measure.
2. They split the roles clearly. IT and OT agree on who owns which part of the data foundation, whether that is, for example OPC-UA servers on site or the central data platform, and how it is kept secure. The engineers on the floor own the models and the decisions they support. Neither has to become the other.
3. They make the process experts visible. When an engineer finds the cause of a problem that has been bothering the plant for years, that story spreads. It builds trust in the technology faster than any steering committee can.

Collaboration doesn’t mean everyone needs to understand everything. It means everyone understands their part, and trusts the other side to do theirs.

Bridging the gap

Bridging the IT/OT divide is essential for succeeding with AI projects. By fostering collaboration and choosing Industrial AI tools that are easy to use, businesses can make sure both IT and OT perspectives are part of the process from the start. Recognizing the importance of that cooperation is key to unlocking the full potential of AI in the process industry, reducing project failure rates and achieving smooth integration across IT and OT.

At Intelecy, our commitment is to provide a user-friendly Industrial AI SaaS solution that improves operations, drives meaningful change, achieves sustainability goals and leaves a lasting impact. The Intelecy Industrial AI Platform lets process engineers and operators harness the power of AI on their own process data, and today they can also work with their models and data directly from AI assistants like Claude and Copilot through our MCP server. Explore real use cases from clients in the process industry who use Intelecy to achieve substantial business value.

Please reach out if you’d like a tailored demo, have questions, or want to discuss whether modern technology really can help close the IT/OT gap. We’re here to help you kick-start and succeed with your AI initiatives!