Summary of Key Points
In the first half of 2026, national consumer associations received nearly one million complaints, with after-sales service issues accounting for the highest proportion (26.79%). New problems arising from AI customer service (such as difficulty in transferring to human representatives, unfulfilled promises, and inaccurate information) have become a growing source of complaints. The main reason companies are keen on using AI customer service is the cost savings: while the average monthly cost for human customer service is 3,000 yuan, the cost for AI is only 99 yuan, which can significantly reduce labor expenses. However, companies often design AI systems as an "invisible filter" to discourage consumers through complex processes. The new national standards implemented on September 1st aim to address these issues, but their effectiveness remains to be seen since they are merely advisory in nature. In this battle of AI customer services, consumers are forced to learn "anti-AI strategies," and while companies may seem to save costs, they could potentially lose customer trust and long-term value.
Detailed Analysis
1. The Secret Behind the Popularity of AI Customer Service: The Cost Savings
Why has AI customer service quickly gained market traction? The key lies in cost savings. The difference is clear: human customer service costs at least 3,000 yuan per month (including social security and training), while AI customer service costs only 99 yuan per month, with each consultation costing between 0.07 and 0.21 yuan—4% to 6% of the human cost. Large e-commerce platforms can save over one million yuan annually, and smaller businesses can reduce costs by 65% (for example, a home goods e-commerce company reduced its monthly cost from 35,000 yuan to 12,000 yuan after implementing AI).
However, companies' "smartness" doesn't stop there: AI can handle 70%-85% of standardized issues (such as checking orders and asking about policies), but they deliberately pretend not to understand complex complaints (like a broken lock that requires immediate attention) by using repetitive scripts or hiding human contact options, leading consumers to give up. Each consumer who gives up seeking relief helps the company save on complaint handling fees. This isn't due to technical limitations; it's a deliberate design to filter out more demanding customers, showing lower complaint numbers in reports and thus lower costs, although customer churn is an implicit consequence.
2. Three Major Traps for Consumers with AI Customer Service
The pitfalls created by AI customer service can be summarized in three points:
- Difficulty in Transferring to Human Representatives: Repeated requests to transfer to a human representative are often ineffective; you must specifically mention "complaints to regulatory authorities" or "12315" to get through (for example, on a certain video platform, you may need to call four times and wait six minutes).
- Unfulfilled AI Promises: Second-hand car platforms may claim to cover transfer fees, but once customers pay, the seller might say the AI response doesn't represent the company's policy—similar to a restaurant screen showing a special offer of 28 yuan, only for the waiter to say the price displayed on the screen doesn't apply.
- Inaccurate Information Provided by AI: AI responses are based on probability rather than actual knowledge. For example, if you ask about whether fees are included, the AI might say "yes" simply because that answer is more common, not because it's a genuine promise.
The essence of these traps is the information asymmetry: consumers assume AI represents the company, but companies use AI as a way to shift responsibility.
3. The New National Standards: Can They End the Hell of Transferring to Human Representatives?
The new standards, implemented on September 1st, set clear rules for AI customer services:
- The option to transfer to a human representative must be clearly displayed and not hidden.
- In five situations, the system must automatically transfer the call to a human representative: too many interactions, user refusal of AI assistance, emergency situations (e.g., a baby needing milk with the door locked), risky scenarios, or delayed AI responses.
- Elderly and disabled individuals should be given priority in being connected to a human representative.
However, these standards are advisory and not mandatory, so their enforcement depends on industry self-regulation and consumer association supervision. They only address whether the transfer is possible; they don't guarantee that issues will be resolved once the call is made to a human representative, which still depends on the company's conduct.
Consumer associations have already taken action, planning special inspections of AI customer services in the second half of 2026, focusing on the transfer process and fulfillment of promises, as a warning to companies.
4. Consumers' "Anti-AI Strategies": Tactics Developed Out of Need
Facing these obstacles, consumers have developed self-help methods:
- Using Key Words: Avoiding the phrase "transfer to human representative" and directly mentioning "complaint to 12315" or "contact regulatory authorities" can trigger a risk alert, allowing for a quicker transfer.
- Timing Tactics: Calling in the middle of the night or early in the morning when the AI system is less active may increase the chances of getting through to a human representative.
- Evidence Preservation: Taking screenshots of AI promises (e.g., records showing a commitment to cover transfer fees) can be used as evidence in case of disputes, as consumer associations recognize these as the company's responsibilities.
- The Ultimate Weapon: Directly contacting 12315 or 12345 to bypass the AI system and seek official intervention.
The popularity of these strategies highlights the irony of the service industry: resolving issues with a real person now requires tactics akin to those used in spy movies.
5. The Dangers of Viewing Consumers as Costs
While companies may initially save money by using AI, the long-term consequences can be detrimental:
- Loss of Trust: A negative experience with AI customer service can make a consumer remember the brand's poor after-sales service and avoid purchasing from them in the future.
- Hidden Customer Churn: The 5% of consumers who are filtered out tend to be more proactive in seeking rights and sensitive to prices, potentially spreading their negative experiences, leading to further customer loss.
- Higher Long-Term Costs: Industry analysts predict that by 2028, the hidden costs associated with poor customer experiences will exceed the savings from using AI. The money saved through AI may ultimately be lost due to increased customer churn.
The essence of service is an investment, not just a cost. While AI customer service aims to improve efficiency, when it's used to prevent issues rather than solve them, it deviates from its purpose. Technology can be efficient, but service must remain warm and attentive; companies can reduce costs, but they should not treat consumers as mere expenses.
In Conclusion
After September 1st, we hope that transferring to a human representative will no longer be a tedious process but a basic service that consumers can easily access. After all, technology is meant to serve people, not to keep them out.