The 80/20 Problem: When AI Gets Most of It Right

AI is becoming increasingly useful in preparing and reviewing construction documents, specifications, reports, schedules, and preliminary design information.

It can organize complex information, produce initial drafts, identify inconsistencies, and support drawing coordination.

Consider a complicated project where AI produces 80% of an initial product. That may represent considerable time savings. But what happens if the remaining 20% contains critical errors?

An overlooked specification, incorrect dimension, missing code requirement, or poorly defined contractual obligation can create consequences far greater than the time saved preparing the document.

In construction, the final 20% isn’t necessarily a minor detail. It may be where the greatest financial and operational risks reside.

The 80/20 scenario is illustrative, but the underlying concern is real: AI-generated work should be treated as a starting point, not a finished professional deliverable.

Qualified architects, engineers, construction managers, and legal professionals must review and approve work within their respective areas of responsibility.

Who Is Using AI on Your Project—and Who Is Checking the Work?

Another important question is: How are your architects, engineers, contractors, and consultants using AI?

AI may already be supporting design development, cost estimating, subcontractor evaluations, scheduling, document preparation, or contract review.

These applications can deliver meaningful efficiencies. However, owners should understand where AI is being used, how the information is verified, and who remains accountable for the final work product.

Who is responsible when a civil drawing fails to identify key information or when a contractor’s AI-generated estimate overlooks a critical scope item?

Owners should consider establishing clear expectations with project partners regarding AI use, including appropriate disclosure, professional review, confidentiality protections, and contractual responsibilities.

Technology may change how work is performed, but it should not create uncertainty about who is responsible for the results.

7 Best Practices for AI In Construction:

1. Establish clear boundaries. Identify which activities can benefit from AI and which require independent verification, professional review, and formal approval.
2. Keep accountability with qualified professionals. AI can support analysis and recommendations, but responsibility for design, safety, contractual obligations, budgets, and execution must remain with the appropriate professionals.
3. Know how your project partners are using AI. Establish expectations for architects, engineers, contractors, and consultants, including review procedures, confidentiality, and responsibility for AI-assisted deliverables.
4. Protect confidential project information. Financial models, proprietary designs, contracts, bid information, and workforce data should only be used in AI systems with appropriate security and data-use protection.
5. Verify before acting. AI-generated schedules, estimates, specifications, and technical recommendations should be checked against reliable project information, applicable requirements, and actual field conditions.
6. Start with measurable applications. Focus on areas where AI can demonstrate value, such as document review, schedule analysis, risk identification, and project reporting. Measure results before expanding its use.
7. Build oversight into project management. Establish who reviews AI outputs, how important decisions are documented, and when issues must be escalated. AI should strengthen existing project controls, not operate outside them.