Brothink wants you to know clearly where there is AI and where there is a person. This page lists the features that use AI, how they are reviewed and the limits you should know about.
1. AI models Brothink uses
Brothink picks a model for each job and combines several providers: Claude (Anthropic), ChatGPT (OpenAI), Gemini (Google) and open-source models run through Hermes or OpenClaw. For models used through business APIs, under the providers' terms, data is not used to train models by default. Ready Agent runs on your own Claude or ChatGPT account; Open Agent can run on infrastructure you control. The AI advisor on the homepage currently uses Claude.
Image generation and video processing may need other specialized models. When such a feature opens, the provider's name is stated on the tool page and in the Data policy.
2. Which features use AI
| Product | What AI does |
|---|---|
| Market Radar | Gathers public information about competitors and sorts customer reviews and comments. |
| Content Studio | Drafts the posting calendar and content in your brand voice. Product photos and video processing are AI features. |
| Sales Trainer | Plays the customer in practice sessions, scores and gives feedback on each turn. |
| Event Games | AI is not used to decide outcomes. Spin results are decided by the server within the prize limits. |
| SuyLuan | Challenges a decision from several angles, scores the evidence and lists what is still unknown. |
| Ops Cockpit | Summarizes operating figures into a morning briefing and suggests items to approve. |
| CSKH Agent | Drafts replies to customers based on real price and stock data. Staff read them and press send. |
3. Human review
AI drafts and suggests, a person approves. Posts, messages and emails sent outside, any spending and deliverables handed to clients all go through a person first. Details in the AI safety commitments.
4. Labeling and self-identification
- Chatbots identify themselves as AI at the start of each conversation.
- AI-generated content is labeled so reviewers and readers know.
5. Limits of AI
- AI can be wrong. It can misread a question, give outdated information or state something untrue as fact.
- Prices, stock and revenue must come from your real data. When there is no data, the tool says it is unknown instead of guessing.
- Results do not replace the advice of a lawyer, accountant or financial professional. SuyLuan is not investment advice.
- Read the results again before using them with customers or for important decisions.
6. Reporting problems
If a result is wrong or harmful, tell Brothink at support@brothink.com or call (+84) 39 405 1006. Include the tool name and when you used it.
