Your AEC Firm Has a Memory Problem. Here Is How to Fix It

Key Takeaways

• UK contractor spent 2 years cleaning 1M+ records for AI readiness
• Misplaced information causes 5% of project costs, ~$5.5M on $109M builds
• Metadata, automated classification, and permissioning are core data principles
• Centralized platforms let AI operate on data in‑place, reducing security risk
• Egnyte’s Project Hub adds ROT management to prepare data for AI

Pulse Analysis

The construction sector’s rush to adopt generative AI has exposed a fundamental data foundation gap. Projects generate terabytes of drawings, contracts, emails and cost models across siloed systems, yet firms often restart each new job without reusable knowledge. Studies presented at AI in AEC 2026 reveal that rework—driven by missing or inconsistent information—eats roughly 5% of total spend, translating to millions of dollars on large‑scale builds. This inefficiency not only erodes margins but also jeopardizes compliance mandates such as the UK Building Safety Act, which requires secure, searchable records for three decades.

Addressing the gap starts with three disciplined practices. First, robust metadata tags every document with context—project phase, discipline, date—so files are discoverable beyond cryptic naming conventions. Second, AI‑powered classification automatically labels files as drawings, contracts or invoices, eliminating manual tagging at scale. Third, granular permissioning controls who can view or edit each asset, protecting sensitive designs and financial terms while ensuring the right stakeholders have timely access. Together, these steps transform scattered files into a searchable knowledge graph, enabling AI to surface insights directly where the data resides.

A centralized data platform embodies this approach, allowing AI models to run in‑place without exporting files to multiple point solutions. Egnyte’s AEC‑focused Project Hub combines collaboration, governance and built‑in AI, automating metadata application, classification and ROT (redundant, outdated, trivial) cleanup. By consolidating project assets into a secure cloud, firms gain a single source of truth that satisfies both operational efficiency and regulatory retention requirements. The result is a faster, lower‑cost AI rollout that delivers measurable value—fewer rework incidents, shorter design cycles and stronger compliance posture. Companies that prioritize data infrastructure now will capture the competitive advantage AI promises in the built environment.