The Project Controls Data Lakehouse Is Not the Goal. Better Decisions Are.

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

Many energy organizations have already made the infrastructure investment.

They have a cloud strategy. They have data tools. They may even have a data warehouse or Lakehouse and more dashboards than anyone wants to admit.

Yet in the outage control room, project review, or executive meeting, the same problem keeps showing up: people are still reconciling numbers before they can make decisions.

That is because a Project Controls Data Lakehouse is not the outcome. It is the foundation.

The outcome is a trusted operating environment where schedule, cost, scope, resources, risk, and operational information tell one coherent story, and where leaders can see emerging risk early enough to do something about it.

At Knowledge Relay, we call that One Truth. Not because every question has a single simplistic answer, but because every approved metric, forecast, and decision should be traceable to governed data, common definitions, and transparent business rules.

What is a Project Controls Data Lakehouse?

A Project Controls Data Lakehouse is a governed data environment that integrates the information used to plan, execute, control, and improve capital projects, outages, and work programs.

It typically brings together structured data such as Primavera P6 schedules, SAP financial data, IBM Maximo or Asset Suite Project Controls data, budgets, and resource information, with semi-structured and unstructured sources such as work packages, risk registers, contracts, technical documents, daily logs, lessons learned, status narratives, images, video, and sensor data.

But the architecture does not create trust by itself.

A Lakehouse becomes a true project-controls capability only when it provides:

• Common definitions for schedule, cost, scope, resources, risk, and performance.
• Source-to-target mapping, data lineage, and historical traceability.
• Validation, cleansing, master-data alignment, and governed calculations.
• Security and role-based access appropriate to the operating environment.
• Workflows, metrics, dashboards, forecasts, and decision support that teams use every day.

That last point is where many modernization programs stall. The platform gets built. The operating model does not.

The executive problem is not data volume. It is decision latency.

Fragmented data is often described as an IT issue. In practice, it is an operating and executive issue.

When project controls, finance, operations, supply chain, and outage teams work from different numbers, meetings become reconciliation exercises. Escalations take longer. Confidence erodes. People fall back on spreadsheets and side calculations because the official view arrives too late or does not answer the operational question.

The cost is not merely inefficient reporting. It is decision latency, the time between a real change in performance and a decision that changes the outcome.

In high-consequence energy operations, that gap matters. A schedule variance that is identified after float is consumed is no longer an early warning. A resource conflict discovered in the field is no longer a planning issue. A cost forecast updated after a work package is executed is not a forecast; it is an explanation.

The U.S. Government Accountability Office’s Schedule Assessment Guide makes the same practical point from a scheduling discipline perspective: reliable schedules require complete activities, logical sequencing, resource consideration, a valid critical path, schedule-risk analysis, updates using actual progress, and a maintained baseline. A dashboard cannot compensate for an unreliable schedule or disconnected underlying data.

Why “One Truth” must include Project Controls context

A single source of truth is often misunderstood as a single database.

That is too narrow.

For Project Controls, One Truth means the organization has a governed way to connect the relationship between the plan and the work:

• The schedule reflects the scope and the logic of execution.
• The resource plan reflects availability, qualifications, constraints, and demand.
• The cost forecast reflects progress, commitments, change, and risk.
• The risk register is connected to activities, work packages, assumptions, and mitigation actions.
• The field, outage, and operational signals can be interpreted in the context of plan-versus-actual performance.

When those relationships are disconnected, teams can produce technically accurate reports that still fail to support the decision in front of them.

This is why Knowledge Relay’s approach begins with the systems customers already use and evolves from a Data Depot to a governed Project Controls Data Lakehouse, then to AI-driven analytics and increasingly automated outcomes. Organizations can begin at the stage that reflects their current maturity. They do not need to rebuild everything before gaining value.

For a deeper view of this principle, read our related article: The Foundation Is Not the Finish Line: How “One Truth” Turns AI into Measurable Plant Performance.

The transition from reporting to operating capability

The practical maturity path has three stages.

Stage 1: Connect the data and automate the reporting work

The first step is often a centralized Data Depot that automatically extracts, integrates, reconciles, and preserves historical snapshots from existing project-controls and operational systems.

The immediate outcomes are practical:

• Less manual data collection.
• More consistent calculations.
• Automated readiness, execution, fleet, executive, and management reporting.
• Faster access to actionable information.
• A reliable foundation for what comes next.

The value is not that a spreadsheet disappears. The value is that teams stop spending their highest-value hours manually assembling the past and can spend more time managing the future.

Stage 2: Govern the “One Truth” required for scale and AI

The next stage expands the structured Data Depot into a Project Controls Data Lakehouse that can incorporate documents, narratives, risk information, external inputs, sensors, and other non-tabular sources.

This stage is where governance becomes operational rather than theoretical. Common definitions, quality controls, lineage, security, access controls, and calculation rules become part of how information is delivered and used.

That is particularly important for AI. The National Institute of Standards and Technology’s AI Risk Management Framework is designed to help organizations incorporate trustworthiness into AI design, development, use, and evaluation. Its core functions, Govern, Map, Measure, and Manage, are a useful lens for energy organizations that want AI to be credible with operations leaders, executives, and regulators.

Stage 3: Apply predictive intelligence to the work that matters

Once the data is accurate, integrated, and governed, AI can move beyond generic experimentation.

This is where organizations can use their project-controls environment to assess schedule risk, forecast completion dates, identify critical-path and near-critical-path risk, forecast cost and resource demand, analyze variance drivers, compare fleet performance, and evaluate what-if scenarios.

The objective is not to replace accountable operators, schedulers, planners, cost engineers, or project managers. It is to give them better evidence sooner.

AI-driven recommendations, such as resource leveling options, schedule optimization, risk mitigation alternatives, and budget-schedule alignment, are most valuable when they are explainable, grounded in governed data, and reviewed by the people responsible for execution.

That is why the biggest obstacle to AI success in project controls is usually not the model. It is the readiness of the data, governance, and business context around the model.

Questions an executive team should ask now

A useful way to evaluate a Data Lakehouse initiative is not to ask, “Do we have a platform?” Ask:

1. Can leaders see schedule, cost, scope, resource, risk, and work-execution information from a common governed foundation?
2. Can a project-controls metric be traced to its sources, definitions, business rules, and historical snapshot?
3. Are outage and project teams spending time interpreting risk or wasting time reconciling reports?
4. Can the organization identify a change in risk early enough to change the outcome?
5. Is AI being applied to a defined operational decision with accountable human oversight?

If the answer to several of those questions is no, the opportunity is not simply to add a new dashboard or launch another pilot. It is to build the operating layer that turns the existing data investment into measurable performance improvement.

The Knowledge Relay perspective

Energy utilities do not need a theoretical data strategy. They need an environment that works under real operating conditions.

For over 40 years, Knowledge Relay has helped nuclear and energy organizations connect existing enterprise systems, establish a governed Project Controls One Truth, automate high-value metrics and reporting, and can now apply AI-driven analytics to schedule, cost, resource, risk, and outage decisions.

The end state is straightforward: less time reconciling data, more time deciding; earlier visibility into risk; and a project-controls capability that increasingly helps plan, forecast, optimize, and execute the work.

If your organization has already invested in data infrastructure but has not yet translated it into better outage, project, or fleet decisions, contact Knowledge Relay to discuss your current maturity and the next practical step.

Frequently Asked Questions