Summary of Key Points
This article focuses on the application of AI in the Taoketong (an industry that helps e-commerce platforms with promotions and earns commissions) sector. The main argument is that while AI has currently improved the efficiency of Taoketongs (for example, by making it faster to write product descriptions and create tools), it has not truly changed the shopping experience for users. The future opportunities for Taoketongs lie in using AI to help users solve the decision-making challenges of "what to buy" and "how to choose," rather than simply increasing the speed at which products are promoted. The article also analyzes the differences in AI capabilities between platforms and independent Taoketongs, and proposes three directions that Taoketongs could explore, along with steps for conducting small experiments to test these ideas.
1. AI in the Taoketong Industry: Busy Behind the Scenes, No Change on the Front
Taoketongs are now using AI for various tasks such as writing product descriptions, creating posters, organizing customer service scripts, and analyzing orders. What used to take a team of people can now be done by one person with the help of AI, significantly increasing efficiency. But what about the users? Their experience remains the same: they join groups to browse products, copy codes to access Taobao, and spend hours deciding which product to buy. The only difference is that the number of product descriptions generated per minute has increased—it's like an old machine running faster, but not being replaced with a new one.
Why isn't this considered a major opportunity? In the past, every breakthrough in the Taoketong industry (such as rebate websites, coupon groups, or social agency apps) brought new benefits or experiences to users, such as getting money back after purchases, instant price discounts, or earning money by sharing. Currently, AI only makes it easier for Taoketongs to work; since users don't receive any new advantages, this doesn't lead to industry-wide changes.
2. Why Can Platforms Implement AI Shopping First?
Platforms have the advantage of holding "core data" that is essential for AI to help with shopping decisions. For example, ChatGPT can help users compare products, Google can check local inventory, and Alibaba's Qianwen allows users to search for products in Taobao using natural language. Platforms can take the lead because they need real-time business data to provide useful recommendations: are there any products in stock? Can deliveries be made in the user's area? Can coupons be used together? These pieces of information are only available to platforms, not to Taoketongs, who only have promotional links.
For instance, if you ask AI "What phone can I buy for under 2000 yuan?" the platform's AI can tell you the available models, their prices, and whether coupons can be used. In contrast, a Taokong's AI can only provide a list of links; it doesn't know if the prices have changed or if the products are in stock—this is where platforms have a significant advantage.
3. True AI Shopping Products Must Overcome Five Challenges
Simply providing an AI with a product database doesn't make it a useful shopping assistant. A truly reliable product must overcome five key challenges:
1. Real Data: The recommended products must be in stock, at the correct price, and eligible for coupons. If the recommended products have increased in price, even the best descriptions are useless.
2. User Context: AI needs to understand the user's specific needs. For example, when buying a phone, a user looking to buy one for their parents may prioritize large screens and long battery life, while a gaming enthusiast might focus on performance. A one-size-fits-all approach won't work.
3. Credible Evidence: AI must provide clear reasons for its recommendations and explain who the products are suitable for. For example, when recommending a vacuum cleaner, it should state that it's suitable for flat floors but not for high-traffic areas to build user trust.
4. Complete Buying Process: The recommendation process should include direct access to coupons, the ability to place orders, and tracking of deliveries, without requiring users to copy links and navigate elsewhere.
5. Profit Distribution: If AI takes control of the decision-making process, what value do traditional Taoketongs left? If AI uses content from reviewers, it needs to share profits with them; otherwise, no one will provide honest reviews.
4. Opportunities for Taoketongs
Taoketongs won't be able to compete directly with large platforms as "universal shopping assistants," but they can explore three potential directions:
1. Specialized AI Advisors: Focus on a specific category where users struggle with decision-making (e.g., home appliances or baby products) and provide in-depth guidance. For example, when choosing a mirror for decoration, the AI could ask about the room layout, size, and style before suggesting three options with their pros and cons, saving users an hour of research time.
2. AI + Human Personal Advisors: Let AI handle 80% of the work (information gathering, product filtering, price tracking), while humans handle the more complex decisions (e.g., providing personalized advice when users are unsure). This approach reduces costs and builds trust.
3. Providing Reliable Consumer Reviews: In the future, users will directly ask AI for product reviews. Those who can provide genuine experiences and tips will become valuable "data providers" for AI systems. For example, by recording that a vacuum cleaner is suitable for small apartments but may run out of battery in larger ones, you can earn commissions when AI uses this information.
5. Take Action Now: Conduct a Small Experiment in 30 Days
Don't invest heavily in developing a large platform immediately; start with a small test:
1. Choose a Category: Pick a product category you are familiar with and where users often have doubts (e.g., air purifiers).
2. Identify Issues: Gather 100 real user questions from group discussions or customer service records (e.g., "What air purifier should I choose based on its airflow?")
3. Create Decision-Making Cards: Provide clear information for each product, including the price, suitable scenarios, and any potential drawbacks. Also, state how much commission you will earn from each recommendation to build trust.
4. Let AI Ask Questions First: When a user asks about an air purifier, let AI ask about the area, budget, and whether there are elderly or children in the household, then suggest 2-3 options.
5. Manual Review: Check the accuracy of AI's recommendations (e.g., ensure prices haven't changed).
6. Track Metrics: Monitor user behavior—see if users are willing to provide additional information, how many recommendations are accepted, and the refund rate. A high refund rate indicates that the AI is just promoting products without understanding user needs.
If users are willing to return for more advice, you're on your way to creating a reliable shopping assistant.
Two Stages of Change Brought by AI
The first stage of change involves AI helping Taoketongs sell products more efficiently (which most are currently doing). The second stage will involve AI making purchasing decisions for users. In the future, what will be valuable is not just the ability to use AI to write effective descriptions, but having users who trust you, understand their needs, and have access to reliable data. To succeed in the second stage, Taoketongs need to shift from helping users save money to helping them save time and make fewer mistakes.
(This article represents the personal views of Lao Hu and is for reference only.)