AI Cost Estimation
AI Project Scheduling
AI Risk Assessment & Mitigation
AI Labor Estimation
Work Breakdown Structure
Design for Manufacturing (DFM)
Should Cost Analysis
Price to Win (PtW)
Total Cost of Ownership
Design to Cost
RFI, RFP, RFQ
Mastering Cost Risk with the CRED Model: A New Approach to Managing Uncertainty
Table of Contents
Artificial intelligence is increasingly seen as essential for planning and estimation, but adoption is proving more difficult than many executives anticipated. According to Galorath’s 2025 State of the Industry Report on Cost, Schedule, and Risk, 63% of respondents said they have not integrated AI into their daily workflows. Even among adopters, most efforts remain cautious: just 4% of organizations said they have automated most of their estimation process. The rest are still experimenting, with almost half using AI for only part of their estimates.
This gap between aspiration and execution reflects a deeper challenge. AI is not failing because it lacks potential. It is faltering because leaders underestimate the systemic barriers that prevent it from taking root: fragmented systems, unclear ownership, inconsistent data, and workforce gaps. Unless executives address these foundational issues, AI risks becoming another pilot project that produces headlines but not measurable impact.
Despite rising investment, confidence in estimation is falling short. Only 12% of professionals said they are “very confident” in their accuracy, compared with more than half who remain only “somewhat confident” and 14% who said they are “not confident at all”. This confidence gap underscores the urgency for executives to approach AI adoption intentionally, not reactively.
Industry data highlights the five mistakes most organizational leaders fear and allow to hold them back.
1. Relying on Spreadsheets as a Foundation
Spreadsheets remain the most common tool for project and cost estimation, but they are also the weakest foundation for AI. The report shows that 56% of organizations operate with partially integrated systems, and 11% remain fully siloed. This means the majority of estimation data lives in disconnected spreadsheets, emails, or departmental databases that AI cannot easily reconcile.
For executives, this reliance creates more than inefficiency. It undermines trust. AI applied to fragmented data simply automates flawed assumptions, accelerating rather than reducing risk. These errors ripple across functions: engineering makes trade-offs based on outdated costs, finance allocates budgets with incomplete information, and procurement negotiates contracts that do not align with actual schedules.
The risk is not hypothetical. The report found that 47% of respondents identified inaccurate data in forecasts as a top challenge, and 65% cited fluctuating material costs as a constant source of disruption. Spreadsheets are not equipped to update dynamically as these inputs shift, making them particularly dangerous when paired with automation.
Executive takeaway: Leaders must transition from spreadsheet-based estimation to integrated, auditable systems where AI models can operate on structured data. Without this baseline, the speed of AI will only magnify existing blind spots.
2. Failing to Integrate AI into Core Systems
When organizations deploy AI pilots, many fail to connect them to the systems executives rely on for planning. The report shows that only 33% of organizations have fully integrated their estimation infrastructure, while more than half reported gaps in cross-functional data sharing. When AI outputs remain isolated from ERP, PLM, and financial platforms, they never inform the decisions that matter most.
This lack of integration creates credibility problems. It’s like hiring a highly intelligent individual for a business-critical position and refusing to give them any contextual awareness of the business. They’ll be smart and insightful, but ultimately misaligned and ineffective. Executives see misaligned data between AI pilots and existing reports. Teams revert to old methods when results are inconsistent. Decision cycles slow instead of speeding up, and AI adoption loses momentum.
Integration is not just technical plumbing. It determines whether insights flow across engineering, procurement, finance, and delivery. Without it, AI cannot achieve its purpose of linking cost, schedule, and risk intelligence to business outcomes.
Executive takeaway: Leaders must insist that AI adoption includes integration to enterprise platforms, not just experimentation in silos. Without system-level alignment, the credibility of AI remains fragile.
3. Leaving Ownership Undefined
AI adoption often stalls because responsibility is unclear. The report shows that workforce capabilities are already strained: 26% of respondents cited lack of training as the top barrier to improving cost estimation, while another 21% pointed to data issues. In this environment, leaving ownership of AI scattered across IT, finance, and engineering is a recipe for paralysis.
Executives cannot assume that AI adoption will self-organize. Without a designated and empowered owner, projects stall in endless evaluation cycles. Governance becomes inconsistent, and investments scatter across pilots that never scale. When something goes wrong — such as a model producing an inaccurate estimate — no one is accountable for fixing the process.
Successful organizations take the opposite approach. They appoint and enable clear ownership, often at the program management or estimation leadership level, supported by cross-functional governance. This ensures that AI is tied to measurable business outcomes and that training, data quality, and integration are managed as strategic enablers, not afterthoughts.
Executive takeaway: Define ownership early. AI in estimation cannot be everyone’s responsibility. It must belong to a function or leader with the authority to set priorities, allocate resources, and enforce governance.
4. Ignoring Explainability and Transparency
Executives need more than outputs from AI. They need traceability. They need to understand how those outputs were created. According to the report, confidence in estimation is low across the board, with only 12% of respondents strongly confident in their current capabilities. In this context, black-box AI does not build credibility — it erodes it.
For regulated industries such as defense, aerospace, and healthcare, auditability is not optional. Estimates must be defensible in reviews, compliant with contracting rules, and transparent enough to withstand scrutiny from auditors, boards, or regulators. AI systems that cannot explain their logic simply will not be trusted at scale.
Transparency builds trust, which is essential to facilitating internal adoption. Finance teams want to see cost drivers. Engineers want to validate technical assumptions. Program managers want to know what happens if a risk variable changes. Without explainability, these groups disengage, and AI remains a novelty instead of a core capability.
Executive takeaway: Treat explainability as a requirement, not an enhancement. AI in estimation must provide traceable assumptions and audit-ready outputs to build trust inside and outside the organization. Leaders should also frame AI as an assistant that supports incremental decision-making, not as a black box promising instant answers. This mindset shift makes adoption more realistic and sustainable.
5. Treating AI as a One-Time Fix
The report makes clear that project environments are volatile. Sixty-five percent of respondents cited fluctuating material costs as their top challenge, while 53% pointed to logistics delays and 50% to trade policy shifts. In this environment, AI cannot be treated as a one-time deployment. Static models will quickly fall out of sync with reality.
Yet many organizations approach AI as if it were a software rollout — something to implement and check off the list. This mindset leaves leaders with outdated models and a false sense of security. The longer the system goes without updates, the more irrelevant its outputs become.
The organizations that report success with AI in the survey are those using it for continuous modeling. They update assumptions regularly, run scenario analyses when conditions shift, and treat AI not as a finished tool but as an evolving capability. That practice is why, according to the report, organizations using real-time modeling are 2.5 times more likely to stay within budget than those relying on static estimates.
Executive takeaway: AI adoption is not a project. It is a capability that requires continuous investment, updated inputs, and ongoing governance. Treating it as a one-time fix will guarantee obsolescence.
Moving Forward
AI in estimation is not failing because the technology is weak. It is failing because leaders underestimate the structural and cultural changes needed to use it well. Spreadsheets, siloed systems, unclear ownership, black-box logic, and static mindsets are all preventable mistakes.
The 2025 State of the Industry Report makes clear that confidence in estimation remains fragile, with only 12% of professionals reporting strong confidence in their accuracy. At the same time, 63% of organizations have yet to integrate AI into their workflows, and more than half continue to operate with fragmented systems. These statistics are not just datapoints — they are warning signals that traditional approaches are no longer enough to manage cost volatility, supply chain disruption, and growing regulatory pressure.
Executives who address these five mistakes head-on will create more than technical improvement. They will establish estimation practices that are credible in boardrooms, defensible with regulators, and trusted across engineering, procurement, and finance teams. That credibility is the foundation of resilience. When conditions change, leaders with integrated, transparent, and continuously updated AI capabilities can adapt quickly without losing stakeholder confidence.
The lesson is clear: AI in estimation must be treated as a long-term capability that evolves with the business environment. Organizations that move beyond pilots, ground their adoption in integrated systems, and embed explainability and governance into every output will not only avoid the pitfalls slowing their peers; they will set the pace for their industries.
10 Step Estimation Process Sample Checklist
View our 10 Step Estimating Process Checklist. This checklist should be tuned to the individual company’s needs and suggestions.
Estimating Total Cost of Ownership (TCO)
Find out how you can use Total Cost of Ownership (TCO) model to create an estimate which includes all the costs generated over the useful life of a given application.
Should Cost Analysis
Learn how Should-Cost Analysis can identify savings opportunities and drive cost efficiency in procurement and manufacturing processes.
ROM Estimate: The First Step Towards a Detailed Project Plan
Find out what ROM (rough order of magnitude) estimate is and why is it a crucial element of every project planning cycle.
Software Maintenance Cost
Find out why accurate estimation of software maintenance costs is critical to proper project management, and how it can make up to roughly 75% of the TCO.
