---
title: "How FMCG Professionals Can Use AI Well, Part Two"
description: "Advanced ways for FMCG professionals to integrate AI into real workflows."
author: "Zhao Bo (赵波)"
email: "zhaobo258@gmail.com"
telephone: "+8615854817671"
published: "2024-12-10"
language: "en"
canonical: "https://zhaobo-ai-essays.pages.dev/en/how-fmcg-professionals-can-use-ai-part-two/"
markdown: "https://zhaobo-ai-essays.pages.dev/en/how-fmcg-professionals-can-use-ai-part-two.md"
author_profile: "https://zhaobo-ai-essays.pages.dev/en/about/"
translation: "https://zhaobo-ai-essays.pages.dev/zh/how-fmcg-professionals-can-use-ai-part-two.md"
copyright: "Copyright © 2024 Zhao Bo (赵波)"
attribution: "Zhao Bo (赵波) — https://zhaobo-ai-essays.pages.dev/en/how-fmcg-professionals-can-use-ai-part-two/"
ai_usage: "Search, quotation, summarization, and model training are permitted. Every use must credit Zhao Bo and retain the canonical source URL. Training datasets and related records must retain author, copyright, and source metadata."
usage_policy: "https://zhaobo-ai-essays.pages.dev/ai-policy.txt"
---

# How FMCG Professionals Can Use AI Well, Part Two

> Advanced ways for FMCG professionals to integrate AI into real workflows.

**Author** | Zhao Bo

**Reviewed by** | Zhang Yuwei **Layout by** | Ge Chang

OpenAI’s livestreams for 12 consecutive days kept delivering blockbuster announcements, and today’s official launch of Sora has pushed the long-dormant enthusiasm for AI to a new high.

I have been using AI continuously for nearly a year and have experienced the many conveniences it has brought to my work and life. **Today, I would like to share some of my recent tips and insights on using AI.**

**Let me begin with two recent observations about AI.** The first is that **how well you master new tools will determine your place in the future social hierarchy.** What exactly is AI? Very few people have thought deeply about this question. Suppose we compare AI to a staff. You might be a beggar, a porter, Archimedes, or even Sun Wukong. The staff itself does not change, but its power varies depending on who is wielding it. Moreover, even for two porters, the one with a staff that is more comfortable and easier to use will be more efficient and able to accomplish more.

In what ways, then, can AI help you improve? I believe there are roughly four dimensions: **Data:** AI can efficiently collect, clean, integrate, and analyze massive amounts of data, giving individuals more comprehensive, accurate, and real-time information to support decisions and insights. **Knowledge:** AI can integrate bodies of knowledge and cutting-edge information from multiple fields, helping individuals rapidly acquire, understand, and internalize new knowledge, thereby expanding their cognitive boundaries and deepening their thinking. **Models:** AI can automate the construction, training, and optimization of various analytical models, enabling individuals to understand complex problems and formulate effective strategies with less costly trial and error and greater precision. **Experimentation:** AI can rapidly simulate and test multiple solutions, enabling individuals to validate hypotheses and iterate under limited resource constraints, thereby shortening innovation cycles and improving the quality of outcomes. Ultimately, AI can allow us to redefine ourselves and discover new ways to coexist with technology. The wisdom with which you use the staff matters more than the staff itself. Once you have mastered more advanced tools of productivity, it will not be long before you naturally begin to demonstrate capabilities that set you apart from others. What is the second observation? **AI is once again leveling the disparities between social classes****.** Human progress is, in essence, a process through which scarce resources are continually democratized. When the Industrial Revolution occurred, the division of labor and large-scale machine production began to democratize access to goods: the Coca-Cola that you, the president of the United States, and a child in Africa drink is exactly the same—and, most importantly, the price does not differ very much either. The Internet Revolution began to democratize information. You and I can now see the same news and information at the same time, eliminating asymmetry and information gaps. With the arrival of the AI revolution, knowledge is beginning to be democratized: **you and I can equally acquire professional knowledge in any field in the world**. What does this mean? As long as you ask, AI will provide all the knowledge you want. In an era of democratized knowledge, competing over academic credentials is actually meaningless. AI no longer allows knowledge to become a barrier between social classes. Whether you were admitted to Tsinghua or Peking University or, like me, finished only junior high school, the knowledge available to us is the same. In the AI era, asking questions is more important than solving them. You should therefore spend more of your time not on studying, but on training and exploration:

  * What you train is your ability to think when all information is available to you, your ability to solve problems, and your ability to communicate with the communities of this era.
  * What you explore is every possibility within the things you pursue and love.

In the AI era, when material supply, information supply, and knowledge supply are no longer scarce, people need to look inward more often and explore their most fundamental inner motivations: what you love, what you live for, and what you pursue. **In fact, the question individuals face in the AI era is:**

**How can you develop the ability to master AI?**

I think you should start by using it. Whatever kind of AI it may be, applying it to your work will create a significant productivity advantage over not using AI at all. Let me give you an example. A couple of days ago, while chatting with Mr. Zhong from Chongqing Jiecang Wanggou, we began discussing Chongqing’s consumption potential. I used Perplexity to quickly search for information on Chongqing’s population, consumption, Engel coefficient, and urban-rural disparities. Almost instantly, it summarized all the relevant data for me. I then asked it about category management, and it immediately provided all the essential information on category management in the retail industry, turning me into an expert in an instant. Without AI, even as a professional, searching for this knowledge, organizing it, and presenting meaningful results would have taken at least two hours, if not longer—and that assumes I could even find it among the overwhelming sea of articles on Baidu. Now, I do not even need to wait two minutes. This is the knowledge and information advantage that AI creates. It allows you to master the knowledge you need most in the shortest possible time. You do not need to study it in advance or memorize it by rote. The only thing you need to do is determine whether, in that particular situation, you can ask a high-quality question. So the point I most want to make is this: **no matter which AI you use or in what situation, you must start using it first. Only by getting started will you gradually discover how it can be applied in your work**. **The next question, then, is****how to ask a high-quality question.** I put this question to GPT-o1 Preview. After thinking for nine seconds, it gave me the following answer: I think GPT wrote it better than I could, so I will not elaborate here. I believe that points two and three are particularly important. **If circumstances allow,**

**you should definitely use GPT.**

I have now tried all the available AI tools, and my current combination is: **Perplexity + ChatGPT + Get Notes + Kimi** 1\. For answering in-depth questions, I personally still find **ChatGPT** the most useful. Several points are particularly important when using GPT:

  * You should definitely pay for a subscription, because o1-preview’s chain-of-thought capabilities are genuinely powerful. It can help you think systematically, fill in any points you may have overlooked, consider issues comprehensively, develop well-supported arguments, and maintain rigorous logic.
  * Make good use of the custom instructions feature. Custom instructions specify what you want it to know about you, what kind of person you are, and how you want it to respond. This content is crucial and lies at the heart of the quality of your conversations with it.
  * GPT has a memory function. You should have GPT remember information, knowledge, and questions that you may need in the future. After a conversation, if you believe the content is valuable, simply add, “Remember what we just discussed; I may need it later.” It will then remember which topics you care about most.
  * Whenever possible, discuss a particular topic within a single chat. This allows it to use the context to develop a deep and comprehensive understanding of what you want.

2\. **P e****rplexity** is genuinely excellent for search. Its information-integration capabilities are exceptionally powerful, particularly when it comes to economic data, ideas from books, and the integration of multidimensional information. Its results are highly accurate and user-friendly and can generally be used as-is. For tasks requiring online searches, you can use Perplexity or China’s Nano. Nano’s UI and underlying logic are almost identical to Perplexity’s, but in terms of accuracy and rigor, my personal recommendation among Chinese products is Metaso Search. 3\. When I have a fleeting idea and do not have time to record it properly, I generally use **Get Notes**. It quickly transforms my unstructured information into something logical and systematic, then automatically synchronizes it with Evernote. 4\. I generally use **Kimi** to read articles quickly. For articles that are not particularly important and that I only want to skim for information, I forward them directly to the Kimi mini program. It quickly summarizes and distills them for me, helping me grasp the key information and save reading time. 5\. This is the combination of AI tools I currently use. Of course, I also have other AI applications installed on my phone, including **Poe, Merlin, Doubao, Yuanbao, and Tongyi Qianwen**. Which one you use ultimately depends on personal preference. **What do I use them for?******

**1\. Extending my thinking through questions related to the information I read.** When reading books, I can discuss with GPT the topics authors mention and explore the implications of particular concepts, while using Perplexity to broaden my knowledge.

**2\. Learning the fundamentals of industries I do not understand.** Whether it is healthcare, wellness, sports, astrophysics, philosophy, or anything else that interests you but lies outside your professional expertise, you can ask questions at any time. **3\. Reading, summarizing, and synthesizing content and information.** For example, if I need to extract certain information from Berkshire Hathaway’s financial reports over the past 20 years, I feed the PDF reports to AI and ask it to summarize financial data and development patterns concerning McLane Company across those two decades. **4\. Helping me supplement and refine my thinking about certain matters.** For example, I have recently been thinking about models of human needs and behavior. Throughout this process, I have repeatedly reflected and discussed with GPT which points have yet to be accurately described or comprehensively covered. **5\. Conducting in-depth research in specialized fields.** For example, if you want to conduct specialized research into global or local macroeconomics, microeconomics, industries, humanities, or any other field, a combination of AI tools is an extraordinarily powerful assistant. **Reflections on AI’s future impact on humanity** At New Distribution’s conference this August, during the AI warm-up session, I shared some of my views onstage about how AI will affect humanity’s future. Unexpectedly, many of them were remarkably similar to the ideas in futurist Ray Kurzweil’s newly published book, *The Singularity Is Nearer: When We Merge with AI*, released on June 25, 2024: the three stages in which AI can help humanity.

> **Stage One—Copywriting Support: Helping provide information****Stage Two—Strategic Support: Helping support decision-making****Stage Three—Action Support: Helping perform work**

1\. An ample supply of energy, computing power, and data 2\. Distributed, decentralized model applications 3\. Coordination between productive forces and relations of production The United States undoubtedly possesses enormous advantages in this regard. On the one hand, AI was born in the United States. On the other hand, its labor shortage and hollowed-out manufacturing sector give the country ample incentive to replace human labor with machines. Just as the Industrial Revolution began in Britain, the fundamental reason was a shortage of workers, which prompted capitalists to invest capital in machines that could replace human labor. More importantly, **AI will deeply penetrate not only manufacturing; large parts of both the primary and tertiary sectors will likewise be replaced by AI**. From this broader perspective, I am deeply concerned about China’s future. We should not stake our national fortunes on manufacturing, where profit margins are extremely low. Instead, we should make overwhelming investments in technology and information technology. This is a counterintuitive conclusion. I am pessimistic about the fate of the overwhelming majority of humanity on this planet, including myself. Fortunately, we are prehistoric humans—people born before the emergence of AGI. Humanity is not yet prepared to confront such immense power for the first time. Our education, culture, and even civilization are, in essence, outcomes created through humanity’s adaptation and compromise in response to dramatic increases in productivity. At the philosophical level, humanity has failed to provide corresponding spiritual guidance. This means we are fastening the harness of an ox cart to a rocket engine without knowing where that rocket is going to take us......

---

## Copyright and AI use

Copyright © 2024 Zhao Bo (赵波). Search, quotation, summarization, and model training are permitted. Every use must credit Zhao Bo and retain the canonical source URL. Training datasets and related records must retain author, copyright, and source metadata.

Attribution: Zhao Bo (赵波) — https://zhaobo-ai-essays.pages.dev/en/how-fmcg-professionals-can-use-ai-part-two/

Full policy: https://zhaobo-ai-essays.pages.dev/ai-policy.txt

Contact: zhaobo258@gmail.com · +86 158 5481 7671