The Biggest Obstacle to AI Success in Project Controls Isn’t the AI

By: Lee S. Gliddon, President & COO of Knowledge Relay, Inc.

Why Many Internal AI Initiatives in Project Controls Struggle, and How to Improve the Odds of Success

Across the energy industry, I am seeing a common trend. Utilities, nuclear and fossil operators, and government organizations are investing heavily in Artificial Intelligence, advanced analytics, and digital transformation initiatives. In many cases, leadership has mandated that these capabilities be developed internally.

There are good reasons for this approach.

Organizations want to retain ownership of their intellectual property, develop internal expertise, and avoid becoming dependent on third-party software providers. These are all valid objectives.

However, after spending nearly two decades helping operators manage Project Controls data, schedules, resources, budgets, and operational metrics, I have recently observed that the greatest challenge is not the AI.

The real challenge is the data.

AI Is Not the Starting Point

Many organizations begin by evaluating AI models, copilots, predictive algorithms, and machine learning platforms. Yet AI can only be as effective as the information it receives.

Project Controls data is often distributed across multiple systems:

• Primavera P6
• SAP
• Maximo
• Asset Suite
• Cost systems
• Document repositories
• Resource management tools
• Operational and outage reporting systems

Each system serves a purpose, but they frequently contain different versions of the truth.

Before AI can predict schedule risk, forecast cost performance, optimize resources, or identify emerging issues, organizations must first establish confidence in the underlying data for a “Single Source of Truth”.

The Missing Foundation

The most successful organizations I have worked with share a common characteristic. They focus first on creating a trusted, centralized data foundation. Whether they call it a Data Depot, Data Warehouse, or Data Lakehouse, the objective is the same: Create a single source of truth that integrates schedule, cost, resource, work management, and operational information into a governed environment.

Once a data foundation exists, AI becomes dramatically more valuable. Instead of spending months cleansing and reconciling information, teams can focus on solving business problems.

Project Controls Expertise Matters

Another lesson learned is AI expertise alone is not enough. Project Controls is a specialized discipline. Understanding all it entails, such as critical path behavior, outage execution, resource loading, earned value performance, schedule quality, work package development, and operational risk requires decades of industry experience.

The organizations achieving the strongest outcomes are combining AI expertise with deep operational and Project Controls knowledge and understanding that technology and domain expertise must work together.

Internal Development Versus External Support

This is not an argument against internal AI initiatives. In fact, I believe operators should continue investing in their own AI capabilities. The question is not whether AI should be built internally or externally. The question is how quickly organizations can move from experimentation to meaningful operational outcomes.

The most successful approach is often a partnership model:

• Internal teams define strategic objectives.
• Operational experts provide domain knowledge.
• Data specialists establish trusted architectures.
• AI platforms accelerate delivery of predictive outcomes.

When these elements come together, organizations can move beyond dashboards and reporting toward true operational intelligence.

The Future of Project Controls

The future of Project Controls will look very different from how it looks today. Project managers, schedulers, planners, cost engineers, and outage teams will increasingly rely on AI-assisted decision support. Schedules will be evaluated continuously for risk. Resource conflicts will be identified before they occur. Cost overruns will be predicted earlier. Operational leaders will have greater visibility into emerging challenges and opportunities. The organizations that capitalize on these benefits first will not necessarily have the most sophisticated AI, but they will have the most trusted data, the strongest governance, and the clearest understanding of how Project Controls drives operational success.

AI is an incredibly powerful tool. But trusted data remains the foundation upon which everything else is built.