How enterprise procurement decisions actually get made now — and why the shift changes everything for JE Dunn.
To measure the gap precisely, we simulated a complete procurement committee for a fictional $2.4B semiconductor fab project.
NovaTech Semiconductor — a fictional company planning a $2.4 billion advanced semiconductor fabrication facility in the American Southwest. The scenario was calibrated to match real-world fab projects announced under CHIPS Act incentives.
Seven stakeholders spanning executive, technical, and commercial/risk functions — the same structure found in real semiconductor procurement committees. Each stakeholder was given 100 questions reflecting their actual decision-making process.
All questions were processed from the perspective of a neutral market analyst — not as JE Dunn asking about itself. This mirrors how a procurement team member would actually query an AI: asking for objective market intelligence, not vendor self-assessment.
700 AI-generated responses revealing exactly how LLMs represent JE Dunn to each member of the buying committee — the recommendations, the caveats, the competitor mentions, and the gaps that determine whether JE Dunn makes the shortlist.
Enterprise construction procurement has undergone two fundamental shifts in the last decade. JE Dunn dominated the first era. The third is where the problem lives.
Pre-2015
Decisions were made through personal networks, industry events, and direct referrals. The GC with the strongest relationships won the shortlist. JE Dunn excelled here — deep client relationships, multi-generational trust, and a reputation built on handshakes and repeat performance.
2015–2023
Google search, trade publication websites, and LinkedIn became part of the buyer journey. Relationships still mattered, but procurement teams began validating GC claims through digital research. JE Dunn maintained its position — the website was professional, the ENR rankings were strong, and the track record was visible.
2024–Present
Buyers now ask ChatGPT, Perplexity, and Claude for GC recommendations before forming a shortlist. The AI's answer — shaped by training data, source authority, and content availability — becomes the first filter. If JE Dunn isn't in the AI's response, they're not in the conversation.
When a semiconductor CFO asks an LLM "Who are the best GCs for a $2B fab project?", the model assembles its answer from a hierarchy of information sources.
ENR rankings, trade publication features, analyst reports, industry association publications. These carry the highest weight because they represent independent validation. Turner has 15+ unique third-party domains citing their work. JE Dunn has 8.
Conference presentations, industry awards, executive thought leadership, co-authored research. These establish credibility through association. Turner's executives appear on 3× more industry panels than JE Dunn's.
Company website, case studies, project pages, press releases. Important but discounted by LLMs as self-promotional. 36.1% of all JE Dunn citations point back to jedunn.com — a sign of first-party dependence.
Revenue figures, employee count, geographic reach, project backlog. These establish scale but not differentiation. JE Dunn is a $7B+ firm ranked #97 on Forbes Private Companies — but this data is rarely surfaced in AI responses about semiconductor capabilities.