The C-suite has embraced a comfortable narrative about artificial intelligence: add AI tools to existing processes, upskill teams on new platforms, create an innovation lab, and transformation will follow. According to\u00a0Tianqiao\u00a0Chen, one of Asia’s most influential technology pioneers and philanthropists, this narrative is not just incomplete. It is actively dangerous.<\/p>\n
In his latest analysis, Chen diagnoses what he calls the “skeuomorphic trap,” the corporate habit of using revolutionary technology to mimic outdated forms. Just as early automobiles were designed to look like horseless carriages, today’s “AI transformed” enterprises are\u00a0largely legacy\u00a0organizations with a chatbot grafted on.<\/p>\n
The result, Chen argues, is not transformation but expensive stagnation.<\/p>\n
The Enable Trap: Addition Logic vs. Multiplication Logic\u00a0<\/strong><\/p>\n Most enterprises today\u00a0operate\u00a0under what Chen calls “addition logic”: Old Process + AI Plugin = New Process. The math seems sound. Productivity rises. Costs drop marginally. Dashboards fill with green arrows.<\/p>\n But structurally, nothing has changed. Humans\u00a0remain\u00a0the “CPU” of the operation, the central processors making judgments, connecting workflows, and transmitting institutional knowledge. AI serves as a peripheral GPU, accelerating calculation but never fundamentally redesigning the system.<\/p>\n Chen uses a visceral metaphor. “This is like strapping an internal combustion engine onto a horse carriage. The speed increases, but can a chassis\u00a0designed\u00a0for the pace of a horse withstand such thrust?”<\/p>\n The answer, for most legacy enterprises, is no. Coordination costs multiply. Integration friction escalates. The organization experiences AI not as a multiplier but as a new layer of complexity grafted onto an already brittle structure.<\/p>\n This is the “AI Enable” phase, and Chen believes it is a\u00a0cul\u00a0de sac that most companies mistake for a highway.<\/p>\n The Native Threshold: Three Mutations and a Phase Change\u00a0<\/strong><\/p>\n To cross from Enabled to Native, Chen\u00a0identifies\u00a0three technical and organizational mutations that must complete simultaneously.<\/p>\n Mutation One: From Probabilistic Fitting to Logical Reasoning\u00a0<\/strong><\/p>\n Traditional large language models predict the next plausible token, a form of\u00a0high speed\u00a0linguistic intuition. They mimic understanding without\u00a0possessing\u00a0it. Newer models, exemplified by systems like OpenAI’s o1, unfold chains of thought internally, evaluating candidate paths and\u00a0self correcting\u00a0before output.<\/p>\n This shift has profound implications for management. In domains with clear rules and boundaries, human roles collapse from “line by line review” to “exception monitoring.” Sampling replaces supervision. The human CPU begins moving out of the critical path.<\/p>\n Mutation Two: From Text Dialogue to Tool Action\u00a0<\/strong><\/p>\n Early AI lived between an input box and an output box, generating advice but requiring humans to execute. Agentic AI\u00a0operates\u00a0APIs, triggers workflows, and executes multi step processes autonomously.<\/p>\n Most\u00a0white collar\u00a0work, Chen\u00a0observes, is simply “moving information from System A to System B according to a rule book.” These rule books, SOPs, compliance checklists, and onboarding manuals, are precisely what agents can operationalize. Humans retreat upstream to strategy or downstream to edge case resolution.<\/p>\n Mutation Three: From Statelessness to Long Term Memory\u00a0<\/strong><\/p>\n Traditional AI forgets. Every interaction is a reboot. Institutional knowledge, the unwritten rules, the lessons from failed projects, lives in the heads of tenured employees, not in infrastructure.<\/p>\n With vector memory, retrieval augmented generation, and long context windows, that constraint is dissolving. Systems can now accumulate years of decisions,\u00a0distill\u00a0patterns from failure, and internalize tacit knowledge into what Chen calls an “enterprise grade hippocampus.”<\/p>\n Experience stops being a human asset and becomes a compounding system asset.<\/p>\n The Melting Point: From Ice Block to Liquid Flow\u00a0<\/strong><\/p>\n When these three mutations align, Chen predicts a structural phase change in the enterprise itself.<\/p>\n Classic management theory, from Taylor to Porter, assumed high coordination costs between organizational units. The value chain was carved into departments, each with defined boundaries and reporting hierarchies. The organization was a crystal, rigid and predictable.<\/p>\n But AI, Chen argues, acts as a thermal event. When agents reduce information friction to near zero, that crystal melts. Work no longer requires routing through fixed\u00a0departments. Resources, talent, and data flow dynamically to wherever the model\u00a0determines\u00a0value is highest.<\/p>\n Chen calls this the “liquefaction of business,” and he believes it\u00a0represents\u00a0the true frontier of AI transformation, not better chatbots, but fundamentally different organizational physics.<\/p>\n Three Diagnostic Questions for the C-Suite\u00a0<\/strong><\/p>\n For executives assessing their position on this continuum, Chen offers three tests:<\/p>\n Few organizations, Chen suggests, pass all three tests.<\/p>\n
\n\n