# How to Tell Whether an AI System Is Real AI or Just AI Washing

Canonical URL: https://oneflash.hk/en/blog/real-ai-vs-fake-ai-system-hong-kong
Language: en-HK
Published: 2026-07-13
Last updated: 2026-07-09

## Direct Answer

For Hong Kong SMEs, the practical way to tell whether an AI system is real or just AI packaging is not to judge how fluent the demo answer sounds. The more important test is whether the AI can read the right business data within controlled permissions, understand workflow state, prepare the next step, create tasks or draft content, hand sensitive decisions back to staff for approval, and leave a clear activity record.

If a system only adds a chat interface, cannot explain its data sources, has no role-based permissions, no approval flow, no audit trail, or cannot connect to real CRM, ERP, WhatsApp, forms or email workflows outside the demo, treat it carefully. It may be AI packaging rather than a practical AI-ready business system.

## What Should Real AI Show?

- Clear data sources: what the AI reads and how fresh the data is.
- Clear permissions: role and action limits for people and AI.
- Workflow state: where an enquiry, customer, order or task currently stands.
- Task support: summarizing, drafting, creating tasks or preparing the next step.
- Human approval: sensitive actions go back to staff.
- Audit trail: AI suggestions, human edits and final versions are recorded.
- Real pilot validation: the system can be tested on a real workflow, not only a preset demo.

## What Should a Vendor Demo Show?

Ask the vendor to demonstrate a realistic, low-risk workflow: a new enquiry arrives from a website form or WhatsApp, the system creates a CRM record, AI summarizes the need and missing information, AI drafts the next reply or creates a follow-up task, staff review and approve, and the system records the final version and next owner.

If the vendor can only show a polished chatbot answer, but cannot show data, tasks, approvals and records, the system may not be ready for business operations.

## AI Claims That Need Caution

- The vendor promises full automation from the beginning.
- The vendor says AI can learn all company data but cannot define boundaries.
- The demo cannot handle exceptions.
- There is no audit trail or activity record.
- The price is low but scope is vague, with no clarity on data preparation, permissions, training, UAT or maintenance.

## oneflash View

oneflash does not treat an AI system as a single chatbot or a standalone AI agent. For Hong Kong SMEs, the more important layer is the business system foundation where people and AI can work together: data, workflow, permissions, approval, audit trail, CRM, ERP, WhatsApp, forms, email and task handoff.

The value of AI is not to perform on the side. It is to help people prepare work inside controlled workflows.

## FAQ

### What is fake AI or AI washing?
AI washing means a company claims a product or service uses AI, but it does not, or the role of AI is exaggerated. In system buying, fake AI can also mean a product only adds an AI label or chat interface but cannot connect to real data, permissions, approvals and workflows.

### Does using ChatGPT or another AI API make a system real AI?
Not necessarily. Using an AI API only means the system has access to model capability. A practical AI system also needs correct data access, workflow state, tool use, human approval and activity records.

### Can an AI agent directly change customer records or send WhatsApp messages?
Some systems can technically do that, but it should not be open from day one. Customer commitments, prices, discounts, payments, complaints, deletions and sensitive data should require staff approval and records.

### What should the first AI pilot be for a Hong Kong SME?
Choose a repetitive, lower-risk workflow with clear data and human review, such as enquiry classification, quotation draft preparation, follow-up reminders, internal knowledge search or order exception alerts.

## Sources

- Cornell Wex: https://www.law.cornell.edu/wex/ai_washing
- SEC: https://www.sec.gov/newsroom/press-releases/2024-36
- FTC: https://www.ftc.gov/business-guidance/blog/2023/02/keep-your-ai-claims-check
- OpenAI: https://cdn.openai.com/business-guides-and-resources/a-practical-guide-to-building-agents.pdf
- OpenAI Agents SDK: https://developers.openai.com/api/docs/guides/agents