Defining AI Commerce
Introducing the Haiwai AI Commerce Dictionary — why terminology has become infrastructure in the AI era
As AI platforms become more sophisticated, they naturally develop their own concepts, terminology, and operating models. Without a common language, documentation becomes inconsistent, product communication becomes confusing, and even AI systems may generate different explanations for the same concept. The Haiwai AI Commerce Dictionary was created as the official terminology reference for the Haiwai ecosystem.
Every genuinely independent technical system eventually grows its own language.
This is not a stylistic choice; it is a structural outcome. Once a platform starts handling objects that did not exist in the search-engine era — agents that reason, knowledge packs that can be retrieved, leads settled on outcomes, attribution split across networks — the old vocabulary runs out. People either borrow an approximate old word or improvise a new one. Both routes lead to the same cost a few months later: one thing, five names.
1. Why every AI platform creates its own language
In the previous era of internet advertising, language was defined by a handful of companies. Impression, click, CPM, CPC, conversion — the meanings were remarkably consistent across the industry, because the delivery model itself was consistent.
The AI era broke that consistency. Each platform organizes "intelligence" differently: some organize it as conversation, some as retrieval, some as delegated execution. Different organization means different abstractions, and different abstractions inevitably fork the vocabulary.
So the same capability — letting an AI answer on behalf of a business — is variously called an assistant, a copilot, an agent, a bot or a concierge. The words look interchangeable. The capability boundaries, liability and billing models behind them are not.
The real cost of forked terminology
When one concept has several names, engineering work is duplicated, documents contradict each other, partners misread commercial terms, and AI systems retrieve competing explanations — producing inconsistent answers to the same question.
2. Why documentation needs a single source of truth
In traditional software engineering, documentation describes the product. In an AI platform, documentation is part of the product, because it is read by humans and models alike.
This is badly underrated. When an agent needs to explain a concept, when a developer needs to implement an interface, when a partner needs to understand a revenue-share agreement, they are all retrieving the same body of text. If that text describes "Merchant Agent" in three ways, the system has three possible behaviours.
A single source of truth is therefore no longer a documentation-style question. It is a runtime correctness question. Terminology consistency directly determines output consistency.
- Humans read the docs → decide how to build
- Agents read the docs → decide how to answer
- Partners read the docs → decide how to contract
- All three paths share one set of definitions; any divergence is amplified three times over
3. Building the Haiwai AI Commerce Dictionary
The Haiwai AI Commerce Dictionary was deliberately not built as "a glossary document." A glossary is static text; a dictionary is a referenced structure.
Every entry is defined as an object that can be retrieved, cited and resolved across languages: a unique canonical name, an explicit category, definitions in Simplified Chinese, Traditional Chinese and English, plus a set of aliases and abbreviations that fold historical usage back into the canonical entry.
The convergence rule matters most. When a term is deprecated it is not deleted — it is redirected to the canonical entry. Historical documents stay readable, while new documents may use only the canonical name.
Entry structure
| Field | Purpose |
|---|---|
| Canonical name | Uniquely identifies a concept; the mandatory form in documentation |
| Category | Places the term in a domain: product, identity, network, advertising, runtime |
| Tri-language definition | Simplified Chinese is authoritative; Traditional and English derive from it |
| Aliases and abbreviations | Resolve colloquial and historical usage back to the canonical entry |
| Related terms | Let readers move through concept structure rather than alphabetical order |
4. How it supports documentation, AI agents, developers, partners and merchants
The value of a dictionary is not how often it is read, but how many systems treat it as authoritative.
- Documentation: every SSOT, handbook and API reference cites one set of definitions, so cross-references stop drifting.
- AI agents: platform concepts are explained from the canonical entry, so the same question does not yield different answers.
- Developers: field names, interface names and business concepts share one vocabulary, cutting translation overhead.
- Partners: the boundaries of settlement, attribution and network roles are fixed in writing, so contracts and implementation agree.
- Merchants: operator-facing copy and internal terminology share one source, lowering the cost of understanding.
“In an AI platform, terminology is not an accessory to documentation. It is an interface of the system.”
5. Looking ahead
We expect a clear trend over the next few years: mature AI platforms will publish their terminology the way they publish APIs, and will version it as part of their external contract.
The reason is straightforward. As more interactions are carried out by agents, interoperability between platforms will depend on whether concepts can be translated accurately. Accurate translation presupposes that each side already holds clear, stable, citable definitions.
The Haiwai AI Commerce Dictionary is the starting point of that view, not its conclusion. Entries will keep expanding as platform capabilities grow, but the rule will not change: one concept, one canonical name, one definition.
Tags
- #AI
- #AI Commerce
- #Documentation
- #Glossary
- #DCCA
- #Agent
- #Platform
- #Terminology
Related articles
Explore the Haiwai AI Commerce Dictionary →
The official terminology reference for the Haiwai ecosystem: canonical names, tri-language definitions, alias resolution and related concepts.