A Consultant's Field Report: Every Corporate AI Project He's Seen Has Failed

A consultant who runs sales and technical delivery for engagements across Fortune 500s and smaller firms argues that organizations are in the grip of AI-driven mania, and that the reality behind the hype is grim. In roughly 18 months his team has watched a 0% success rate on AI initiatives — not just projects they were hired to fix, but ones observed in passing. He contends that the failures are systemic: companies are already bad at shipping software, and AI projects inherit every normal failure mode plus new ones tied to the technology’s novelty and the sheer scale of misdirected investment.

The most common flops are internal and customer-facing chatbots. Internal bots see little adoption because an LLM can only surface what an organization has actually documented, and corporate documentation is thin. Customer bots deliver poor experiences while quietly hiding their own failures — he cites a Mitsubishi voice agent that smoothly promised a callback after his car broke down and then never delivered one, an outcome that likely registered as a resolved case in the metrics even as it cost the company a sale. He argues that project leaders systematically avoid tracking whether tools are used at all, or track metrics that are easy to game, which is why nearly every corporate claim of “massive AI productivity gains” is, in his telling, false.

The deeper problem is cultural. Executives who admit doubt get pushed out and skeptical employees get “randomly” laid off, so every actor — boards, vendors, consultants, staff — is incentivized to misrepresent results. In firms of 500-plus employees, he says, career advancement and even continued employment increasingly hinge on repeated, near-religious professions of faith in AI’s transformative power. The essay’s stated purpose is less to persuade the true believers than to reassure the frustrated, fearful people surrounded by this dynamic that they are not the crazy ones.