If you asked a carriage driver in the late 19th century what he needed most, he would almost certainly have said, “I need fast horses.” He would never have asked for an internal combustion engine.
In a recent thought-provoking analysis, Tianqiao Chen, founder of Shanda Group and Tianqiao and Chrissy Chen Institute (TCCI), uses this classic metaphor to diagnose a major mistake in modern corporate strategy. He argues that today's business leaders are falling into the “skeuomorphic pitfall” of imitating old world shapes rather than leveraging AI to create something truly new.
According to Chen, most companies simply become “AI-enabled” by adding an AI button to legacy software or porting an AI department to an outdated hierarchy. Although this approach is comfortable in the short term, Chen warns that it ultimately leads to a “dead end” that extends the lifespan of systems that deserve to become obsolete. A true revolution, he argues, requires “recoding business from the genetic level.”
Chen's framework outlines three distinct evolutionary stages: AI Enable, AI Native, and AI Awaken.
Stage I: Enabling AI — The Additive Logic Trap
Chen explains that the current “AI enablement” phase relies on a simple additional logic: old process + AI plugin = new process.
In this model, humans are still the “CPU” of the workflow, the central processor that handles decisions and connections, and the AI simply acts as a powerful “GPU” to speed up calculations. Chen compares this to strapping an internal combustion engine to a horse-drawn carriage. As speed increases, the chassis will eventually break apart under thrust.
For organizations to move from this addition phase to a true multiplication phase, Chen identifies three key technological “mutations” that need to occur.
- From probabilistic fitting to logical reasoning: AI moves from predicting the next word to unfolding internal chains of thought, shifting the human role from line-by-line review to “watching for exceptions.”
- From text interactions to tool actions: Agents are evolving from chatbots offering advice to autonomous entities executing complex workflows across APIs and browsers.
- From statelessness to long-term memory: A system that develops an “enterprise-grade hippocampus” that preserves organizational memory, transforming experience from human assets to system assets.
Stage II: AI Native — Business Liquification
Once these mutations are complete, the business world will reach a tipping point. Chen defines the AI native stage as the moment when “AI becomes the CPU” and humans move into strategy and exception management.
He uses the vivid metaphor of melting ice to describe this change. Traditional companies are built like solid blocks of ice, with rigid divisions designed to minimize high connectivity costs. But Chen argues that AI can act as a huge source of heat. Rigid corporate structures “melt” into a fluid state as agents reduce information friction to near zero. Data, people, and resources automatically start flowing where they need to go, without the need for complex management skeletons.
To help leaders determine whether they have reached this stage, Chen suggests three litmus tests.
- Survival question: If AI is removed, will the business simply slow down (enabled) or collapse (native)?
- Flow question: Do AI agents “handshake” and pass tasks directly to each other, or are humans still connecting the nodes?
- Memory Question: Do systems “eat up experience” and automatically translate human errors into new system rules?
Stage III: AI Awakening — Final Boundary
Beyond efficiency is what Chen calls the “AI awakening stage,” a stage in which the fundamental definition of work is questioned.
In this stage, the AI evolves from a “doer” to a “discoverer,” penetrating the wilderness to find laws and solutions never before seen by humans. Mr. Chen faces a tough challenge. As AI begins to question its purpose or rewrite its reward functions, humanity will be faced with a new kind of will.
Why allow this evolution? Chen's answer is pragmatic: “To win.” He argues that the limits for AI-native companies are ultimately the limits of human cognition. To find a breakthrough, organizations may need to allow AI to go beyond human logic and define “what is better.”
Conclusion: A challenge to management
Chen's analysis concludes with a chilling question for today's managers. As you move from enable to native and finally touch wake upwe are dismantling the last outer moat of human intelligence.
“When this silicon species not only works harder than me, but also begins to understand ‘what is right’ better than I do… will there still be a need for my presence?”
For Chen, this transition is inevitable. The question is no longer about which tools to buy, but how to live in a world where correctness is calculated and decision-making is outsourced.
