news

한어Русский языкEnglishFrançaisIndonesianSanskrit日本語DeutschPortuguêsΕλληνικάespañolItalianoSuomalainenLatina

Source: Cover News

Cover News reporter Su Yu and intern Wang Ziyi

"Hello! I'm the delivery guy who delivers food to you. Can you give me a five-star review? The platform can give me a one-dollar reward..." For some time, Zhang Fang, who lives in Chaoyang District, Beijing, has often received such unfamiliar calls. Upon careful listening, it sounds like the "voice of AI (artificial intelligence)."

Review websites with star ratings, must-eat lists, and recommended dishes have gradually become a "consumer guide". As a result, stores have been urging consumers to give good reviews in a variety of ways, such as by sending text messages, making phone calls, and leaving messages, which has also caused a lot of trouble for consumers.

Cover News reporters noticed that on social platforms, topics such as "robots asking for good reviews, AI food delivery boys" were hotly discussed, and many netizens said, "I received a call from the rider asking for good reviews after ordering food, which feels like disguised harassment!"

Who is the deliveryman who asked for good reviews? The reporter found that it was suspected that the "drama" was planned by the merchant and the agency operation company, using AI to generate voice, pretending to be a deliveryman, and using words such as "platform assessment" and "high temperature subsidy" to guide consumers to give five-star reviews to the merchant.

"AI delivery man" has been repeatedly complained about

Customers’ goodwill has become a “rating tool”

"I received the same call twice, and then I realized I was scammed. How could it be possible that the takeaways ordered from different stores had the same voice?" Zhang Fang told reporters that when she asked him which order it was, he kept repeating that it was the most recent order, unable to tell the specific store name, and paused for a few seconds each time he answered, which sounded like an AI voice answer.

Some consumers found that they received "Ai follow-up calls". Screenshot

Some netizens uploaded the audio of the "AI follow-up call". Screenshot

Zhang Fang is not the only one who has had a similar experience. Some people even uploaded their conversations with the "AI food delivery guy" to social platforms, and many netizens posted comments saying that they were also asked for good reviews in the same way. "If they use AI actors to get good reviews for food delivery, how can it be real?"

The reporter found 557 complaints about "asking for good reviews" on the "Black Cat Complaint" consumer service platform, including complaints about "AI harassment calls." Among them, one consumer complained to Meituan, "After ordering takeout, a strange phone call came in, and the content was a robot voice broadcasting that I had to give the rider a good review."

Some consumers have reported this to the Black Cat Complaint Platform. Screenshot

When they learned that someone was "impersonating a rider to ask for good reviews," some consumers posted that "my kindness and compassion were taken advantage of. I originally wanted to get high temperature subsidies for the rider out of goodwill, but I didn't expect it to become a tool to manipulate the ratings."

Takeaway store operation

A "grey industrial chain" has been formed

The reporter found that many companies in the market provide services such as "favorable review follow-up" and "favorable review hosting" to takeaway merchants.

The reporter searched for "takeout delivery agent operation" on a search platform and found more than 30 companies recommended. A staff member of one of the companies told the reporter that agent operation of store reviews can quickly improve store ratings, and stores with high ratings can improve merchant rankings and store conversion rates. "If you increase the rate of positive reviews, the platform will define you as a high-quality merchant, and the platform will promote traffic, and the exposure will be significantly improved, and the number of orders will naturally increase."

"For example, a customer ordered food from other restaurants of the same category before and did not give a good review, but today he ordered food from this store and gave a good review. The system will think that the quality and taste of this store are better than others, and will give priority to pushing this store, thus bringing in more accurate traffic." said the above staff member.

The staff member also provided a relevant case, a store called "Yunnan Copper Pot Rice Noodles", which received 31 orders, 19 calls, and 8 positive reviews on a certain day. "Through 20 days of continuous follow-up visits, the store's score increased from 4 points to 4.7 points."

The takeaway company said that the fees are based on positive reviews. Screenshot

How does the “intelligent return visit robot” cost?

2.5 yuan per piece

Regarding the "return visit" fee, the staff member said that if AI is used, a customer's positive review will cost 2.5 yuan, while manual calls will cost 3.5 yuan. "AI calls use an intelligent return visit robot system, so it is best to ensure the quality of your food, otherwise the customer may give a bad review."

A staff member of another operating company said that the price of positive reviews obtained through AI voice follow-up is 3 yuan per review, and the success rate of positive reviews is about 30%. Merchants do not need to pay for “failed follow-up visits”, and the actual fees collected are charged according to the number of five-star positive reviews obtained.

When the reporter asked whether this type of "favorable follow-up" faced the risk of being reported for violations, the other party said, "Basically not, it's just that the platform keeps having new rules."

The reporter noticed that according to the "Meituan Merchant Evaluation Integrity Management Measures", harassing users to disrupt the evaluation order and calling on users to write false positive reviews are considered hype and interference with platform evaluation, which is a violation. Violations will result in deduction of "integrity points". If merchants are deducted integrity points, the system will trigger punishment measures based on the accumulated deductions, including ranking demotion, prohibition from being on the list, etc.

A takeaway company promotes its "favorable feedback" campaign. Screenshot

How to improve the rate of favorable comments through “return call”

Merchants: One bad review needs multiple good reviews to make up for it

A Meituan food delivery merchant revealed that positive reviews account for a large proportion of the platform's assessment. When the score is lower than 4.8 points, the merchant's traffic will not be very good, which will directly affect the exposure and order volume on the platform.

"If the rating is low, 20 positive reviews will get it to around 4.6, which is relatively quick. If you want to get to 4.9, it may take a week and take a long time."

The reporter also consulted another food delivery agency, whose staff said that after confirming the cooperation, the store will inform the food delivery platform account and password. "We will make a follow-up call based on the order, and provide follow-up statistics the next morning, including the number of orders, number of calls, number of connections, etc."

Why do many customers receive phone calls as soon as they receive their food delivery? He revealed that when entering the merchant system, AI can be used to automatically call customers, and each order will be automatically returned, so many customers can receive phone calls as soon as they receive their food delivery.

In addition to asking for positive reviews from the operating company, store staff will also call customers to seek positive reviews. Li Yuzhou, who works part-time at Wallace, revealed that the calls are mainly made to Meituan users, because a bad review may need to be compensated by many good reviews.

"Hello! Today our store is giving away snacks. Can you give us a five-star rating after receiving the food?" Li Yuzhou usually starts the call in this way. She said that although most users verbally agree, more than 60% of them will not give us a rating.

She also revealed that she needs to ask for 5 good reviews for her store every day. If the store's score is lower than 4.7 points, she needs to make more calls to improve the score. When asked why she didn't find an agent to ask for good reviews, she said, "Our store manager doesn't want to spend the extra money. And our store's score is well controlled. If the score drops, it can basically be raised back by making a phone call in one or two days."

The operating company staff claimed that it could "avoid bad reviews". Screenshot

Why would someone lie and pretend to be a delivery man to ask for a good review?

Operating company: Users will sympathize with the riders

The reporter found that most operating companies asked customers for good reviews in the name of deliverymen, rather than in the name of stores. What is the "trick" here? A staff member of a certain Xiang takeaway operation company said that using the rhetoric of claiming to be a rider, the conversion rate of good reviews can reach more than 30%. "Delivering food on a hot day, users will definitely sympathize with the rider."

The reporter found that consumers' evaluation of riders only involves "satisfaction/dissatisfaction" and specific performance, such as "courtesy and enthusiasm", "wearing work clothes", etc., and only set up a five-star evaluation mechanism for merchants. However, when customers evaluate riders, they must evaluate the store and complete it before submitting. When many consumers received calls from "fake riders" and gave good reviews to the riders, they also "conveniently" gave good reviews to the merchants.

Will the deliverymen actively ask for good reviews from users? A deliveryman told the reporter, "Good reviews are basically useless, but bad reviews are not allowed. A bad review will deduct 3 yuan from the customer's payment."

The customer service of Meituan’s food delivery platform stated that those who make follow-up calls in the identity of riders and ask for good reviews using reasons such as “high temperature subsidies” and “regularization” are generally virtual numbers that merchants use to impersonate riders to call customers.

The customer service also said that the platform does not assess the number of positive reviews for riders, nor does it offer extra rewards. Therefore, there is no need for riders to spend time calling users.

"At present, Meituan is actively rectifying this phenomenon and will take measures such as splitting the evaluation entrance or conducting after-sales intervention on virtual numbers." The customer service representative said.

AI asks for positive reviews

Is there any suspected violation of the law?

Zhu Wei, vice president of the Beijing E-Commerce Law Research Association and member of the Expert Committee of the China Consumers Association, told the Cover News reporter that the "Regulations on the Implementation of the Consumer Protection Law" stipulate that without the consent of the consumer, operators may not send commercial information to consumers or make commercial calls.

Recently, the State Administration for Market Regulation announced the "Interim Provisions on Anti-Unfair Competition on the Internet", which will come into effect on September 1. It clearly states that users shall not be induced to make designated positive comments, likes, targeted votes and other interactive behaviors by means of cashback, red envelopes, cards and coupons. False or misleading commercial propaganda on user reviews shall not be made to deceive or mislead consumers or the relevant public.

"This behavior obviously allows users to create false reviews, violates the wishes of consumers, and is an act of disrupting market competition order and unfair competition." Zhu Wei said that calling users' phones through an agency company also involves infringement of users' personal information rights. The "Network Transaction Supervision and Administration Measures" clearly stipulate that network transaction operators and their staff must strictly keep the collected personal information confidential. Except for cooperating with regulatory and law enforcement activities in accordance with the law, they shall not provide the information to any third party, including related parties, without the authorization and consent of the person whose information is collected.

Regarding AI phone automatic dialing, Zhu Wei said that if the agency operator installs plug-ins and systems on the food delivery platform, it will be a major network security issue, which will undermine the market competition order and the Internet industry ecology. The food delivery platform should conduct an investigation and deal with it through legal and technical levels.

"The relevant platforms must punish false reviews, manage the platform ecology, guide merchants to operate in compliance and compete fairly, and safeguard the legitimate rights and interests of consumers," said Zhu Wei.

(Zhang Fang and Li Yuzhou are pseudonyms in this article)