A senior executive at a billion‑dollar firm rejected a working AI demo because the vendor was small and unfamiliar, preferring a big name for insurance. The demo proved the system could deliver high accuracy while competitors quoted $100,000‑$300,000 for a promise of 99.5% accuracy, yet the buyer still chose risk‑averse certainty.

The post traces a 60‑year pattern of enterprise knowledge‑management failures, citing HP’s $11.1 billion Autonomy purchase and subsequent write‑off. It argues that buyers prize familiarity—vendor name, language stack, and analyst labels—over technical fit, turning each procurement into a safety bet that stifles innovation.

Examples like SharePoint illustrate the gap: 200 million users yet widespread dissatisfaction, driven by the product’s ubiquity rather than its usefulness. The author cites Rich Hickey’s “Simple Made Easy” to frame the problem: simple systems that separate data from logic are easier to adopt, but enterprise decisions reward familiarity, not correctness.

Ultimately, the article warns that the continued preference for familiar vendors will keep enterprises locked into ineffective knowledge systems, costing hundreds of billions in lost productivity and sunk capital.