# AI training for marketing teams in Hong Kong: from research to a finished proposal > Learn how to implement in-house AI training for a Hong Kong marketing team, from research and Copilot agents to understanding what AI should not do. - URL: https://hyperfocusam.com/blog/ai-training-for-marketing-teams-hong-kong/ - Language: en - Author: Sam Wong (黃力桐), https://hyperfocusam.com/ - 中文版: https://hyperfocusam.com/blog/ai-training-for-marketing-teams-hong-kong-tc/ (Markdown: https://hyperfocusam.com/blog/ai-training-for-marketing-teams-hong-kong-tc/index.md) - Published: 2026-10-30; revised 2026-10-30 --- **Short answer:** AI training for a marketing team works best when the team takes one real-looking brief from research to a finished proposal in the session, using the tools the company allows. Along the way it learns where AI is strong (research, ideas, drafts, visuals, prototypes, repeat reports) and where a person must stay in charge (brand judgement, facts, and exact details such as colour matching). *Sam Wong (黃力桐) is a corporate AI trainer based in Hong Kong. 11,000+ people trained, 70+ organisations, 13 countries.* ## What should AI training for a marketing team cover? The chain of work the team already does, end to end: 1. **Research:** market and customer questions, with the rule that every idea needs data behind it. 2. **Ideas and copy:** directions, headlines and variants to choose from. 3. **Visuals:** mood boards and product visuals for a concept. 4. **Pulling it together:** one source pack the whole team can question. 5. **Prototype:** a simple landing page or deck to show the idea. 6. **Repeat work:** weekly reports and summaries handled by a saved agent. Tools change every few months. What lasts is the habit of linking research, ideas, visuals and the pitch into one chain. ## What should marketing not hand to AI? The interviews before a programme usually tell you. At Chow Tai Fook, the team heads listed what AI should not be used for, such as exact CMYK colour matching and gold-weight estimation, so the programme focused on concept development and reporting. Every team has its own list. Facts, claims and the final brand call stay with a person. ## What did this look like at Chow Tai Fook? [Chow Tai Fook's Hong Kong product development teams](https://hyperfocusam.com/en/cases/chow-tai-fook-ai-design-thinking/) have worked with me three times since 2024. For the January 2026 programme I interviewed the heads of all six teams in one day, then mixed design and project-management colleagues into cross-functional groups. In one afternoon each group went from market research to a complete design proposal and pitched it. The groups used ChatGPT or Gemini for research, Midjourney and Nano Banana Pro for mood boards and visuals, NotebookLM to pull the material together, and Lovable for a campaign landing-page prototype. All five groups finished their proposals in the session. ## Can a marketing team build AI agents inside its own Copilot? Yes, and it is often the better route than adding new tools. At [CLP](https://hyperfocusam.com/en/cases/clp-ai-agents-module/), I taught a hands-on module for the marketing, sales and customer-insight teams and their managers. Staff already used enterprise Microsoft 365 Copilot, so the module stayed inside that approved environment. Participants built agents through five cases: a visual translator, data analysis, weekly-report automation, a pitch assistant and a presentation summariser. In the session, they built a visual-translator agent that read a real utility bill. ## How is in-house training different from a public AI marketing course? Public AI marketing courses suit one marketer who wants a broad grounding. In-house training puts the whole team on one brief, in the tools and brand rules it actually works with, and ends with work the team can show. It also brings designers, planners and project managers into the same room, which is where most proposals slow down. ## What should a marketing team prepare? 1. The approved tools, and any brand or data rules. 2. One past brief the team can rework with fictional details. 3. A short list of what AI must not touch in your business. For how I run training for teams across a company, see [corporate AI training in Hong Kong](https://hyperfocusam.com/corporate-ai-training-hong-kong/) ([中文](https://hyperfocusam.com/zh/corporate-ai-training-hong-kong/)). ## FAQ ### What should AI training for a marketing team in Hong Kong cover? AI training for a marketing team should cover the entire workflow, including data-backed research, copy variants, visual creation, and material consolidation. It should involve building prototypes, such as landing pages, and creating saved agents for weekly reports. The most effective training format involves taking a realistic brief from the initial research stage through to a finished proposal within a single session. ### What should a marketing team not use AI for? Marketing teams should not use AI for tasks requiring exact technical details or final human judgement. In Sam Wong's programme for Chow Tai Fook, team heads identified tasks such as exact CMYK colour matching and gold-weight estimation as off-limits. Training should instead focus on concept development and reporting, ensuring that facts, claims, and final brand decisions remain with a human. ### Can marketing teams build AI agents in Microsoft Copilot? Yes, marketing teams can build AI agents within enterprise software. At CLP, the marketing, sales, and customer-insight teams built agents inside Microsoft 365 Copilot during a hands-on module with Sam Wong in December 2025. The five specific use cases included a visual translator, data analysis tool, weekly-report automation, a pitch assistant, and a presentation summariser. ### Which AI tools did Chow Tai Fook's product teams use in training? During the January 2026 programme, Chow Tai Fook product teams used ChatGPT and Gemini for market and customer research. They utilised Midjourney and Nano Banana Pro for mood boards and product visuals, NotebookLM to synthesise materials, and Lovable for campaign landing-page prototypes. All five groups completed a design proposal and delivered a pitch within one afternoon. ### How is in-house AI marketing training different from a public course? While a public course trains individual marketers using general examples, in-house training involves the entire team, including designers, planners, and project managers. It focuses on a specific brief using the company's own tools and brand rules. The training is tailored through interviews with team heads before the session, ensuring it concludes with tangible work that the team can demonstrate. --- **Planning AI training for your marketing team?** [Message me on WhatsApp](https://wa.me/85264315177?text=Hi%20Sam%2C%20I%20read%20your%20article%20about%20AI%20training%20for%20marketing%20teams) and tell me a little about your team. ## Sources - Chow Tai Fook, AI and design thinking for product development teams: [/en/cases/chow-tai-fook-ai-design-thinking/](https://hyperfocusam.com/en/cases/chow-tai-fook-ai-design-thinking/) (read 30 October 2026). - CLP Power, AI-agent module: [/en/cases/clp-ai-agents-module/](https://hyperfocusam.com/en/cases/clp-ai-agents-module/) (read 30 October 2026).