Everything Old Is New Again: Purchased Models, Off-the-Shelf Ontologies, and the Recurring Failure to Model Enterprise Reality
Off-the-shelf ontologies are being positioned as the semantic operating system for banking and other complex industries. They promise shared meaning, interoperability, regulatory traceability, and improved grounding for analytics, automation, and AI. These promises are not without merit. Standardized semantic assets can improve consistency across fragmented environments and can provide a common vocabulary for regulatory and operational use cases.
Sounds catchy, doesn’t it? Inspirational, even. Move fast, break things, learn quickly. But there is a hidden danger behind that mindset – one that quietly conditions teams, leaders, and even entire organizations to accept failure as the destination rather than as a data point along the journey.
Yet the current enthusiasm for purchased ontologies repeats a pattern that has appeared repeatedly for decades under different labels: reference architectures, industry models, enterprise blueprints, and packaged best practices. The core promise remains unchanged: buy a prebuilt model of the enterprise domain, accelerate transformation, and reduce risk. The recurring result also remains largely unchanged: significant expenditures, extended reconciliation efforts, heavy customization, slow adoption, weak executive understanding of value, and eventual disappointment.
