The Irreversible Trend of AI Customer Service Widespread Application: Where Lies the 'Boundary' in Marketing?
-
Though the voice on the other end sounds human, attempts to interrupt or ask questions are ignored as it continues its monologue. Unnoticed, AI calls have replaced traditional business text messages. Yet, when actively seeking customer service, AI becomes a 'roadblock'.
Driven by cost-efficiency considerations, the rise of AI customer service is unstoppable. However, in the bustling wave of business opportunities, there remains a long path to truly earning the label of 'intelligent'.
Frustration
AI Customer Service Becomes a Nuisance
"Hello, this is a premium renovation service tailored for Beijing customers..." "Hello, this is XX customer service. Have you received the delivered package?" "Long time no see! XX platform has sent you a discount coupon—don’t forget to check it out!"
Resident Mr. Yu receives such calls every few days, all from AI customer service. These 'digital employees' operate tirelessly through automated software. Fearing he might miss work-related calls, Mr. Yu answers every one, but the AI calls have become a constant annoyance. "Before, they mostly sent text messages, but recently, direct calls have increased. If it were a real person, I could ask them to stop calling, but with a robot, there’s nothing I can do."
"First, a real agent calls to promote benefits, then switches to AI for further explanation—it feels so敷衍." "Sometimes I’m busy with work or taking a nap, and suddenly a call comes in. When I answer, it’s just an AI, leaving me frustrated with no outlet." On social media, many users share similar complaints. Industries like finance, real estate, logistics, telecommunications, and tourism, which require large-scale customer outreach, are the worst offenders for AI calls.
Meanwhile, when people actually need help and try to contact customer service, they find AI still blocking their way—whether through online platforms or phone calls. It either lists several pre-set questions or plays a string of functions for selection. If the query falls outside these options, communication becomes futile, leaving consumers frustrated with robotic, pre-programmed responses.
Reaching a human agent has become a test of patience. After navigating multiple hurdles and enduring endless 'all agents are busy' hold music, the likely outcome isn’t a warm "Hello" but an automatic disconnect. Online platforms that finally connect users to a human agent often find them too busy for one-on-one responses. If the chat window closes during the wait, the next agent won’t be the same, forcing users to repeat their issues from scratch. After a few rounds, the problem remains unsolved, and the consumer is left exasperated.
Market
Relentless Promotion of 'Digital Employees'
The expanding market for AI in customer service isn’t just a perception—it’s backed by data. A 2023 China Intelligent Customer Service Market Report shows that intelligent customer service software dominates the industry, accounting for about 80% of the market share. In 2022, China’s intelligent customer service market reached 6.68 billion yuan, with projections suggesting it could grow to 18.13 billion yuan by 2027.
Meanwhile, the number of human customer service agents is declining. In the finance sector, a major adopter of AI customer service, data from the China Banking Association’s Customer Service and Remote Banking Development Report reveals that after five consecutive years of growth, the number of customer service and remote banking employees dropped for the first time in 2021 to 50,200, and further plummeted to 43,800 in 2022.
The rise of AI and decline of human agents reflect a clear economic calculation. On recruitment platforms, monthly salaries for phone agents range from 3,000–4,000 yuan to 6,000–7,000 yuan depending on the city, not to mention additional costs for office space, equipment, and training. Replacing humans with AI is a top choice for cost-cutting.
"Standardized scripts, standardized emotions, standardized analysis," "robots simulate human responses to guide sales intelligently," "import customer data with one click, automate bulk calls, and classify customer intent accurately"—sensing market opportunities, tech companies big and small are aggressively promoting their 'digital employees.' Chen Yan (pseudonym), a business manager at a Beijing tech company, remarked, "AI customer service? It’s booming—every company is jumping on the bandwagon!"
Chen Yan explained that the company's clients primarily focus on two types of products. The first is intelligent online customer service, which builds a robot vocabulary database for clients. This involves establishing professional terminology, casual conversation libraries, etc., and later expanding the knowledge base based on customer feedback to improve the robot's response accuracy. Depending on the version, annual software maintenance costs range from 5,000 to 50,000 yuan.
The other type of product, known as "proactive outreach" in the industry, is called intelligent outbound calling. It provides clients with pre-developed systems that incorporate real human recordings and scripts. Once the backend processes are set up, the system can automatically make calls through virtual operator numbers for follow-ups, notifications, and other services. During calls, the system can also capture keywords to tag customers with different "labels" and prioritize them, with annual maintenance costs as low as 3,000 to 4,000 yuan.
Compared to the annual salary of a human employee, AI's "wage" of just a few thousand to tens of thousands of yuan is a clear winner. Moreover, AI customer service can operate 24/7, responding to inquiries at any time. During marketing campaigns, it remains unaffected by complaints or fatigue, delivering extremely high efficiency. CCTV's 3·15 Gala once revealed that some companies developed AI systems for spam calls, with robots capable of making 5,000 calls a day—far surpassing the 200–300 calls made by humans. One company even made over 4 billion spam calls in a year.
Development
The Trend of Widespread Application Is Irreversible
To reduce AI call harassment while avoiding missing important calls, many users opt for voice assistants to answer calls and transcribe recordings into text. When AI encounters AI, they sometimes engage in amusing conversations, proving the adage "one thing subdues another."
"In the past, marketing calls required hiring people to make them. With technological advancements, intelligent outbound calling has become very common," said Liu Xingliang, a digital economy scholar, reflecting on the growing prevalence of AI customer service. "Sometimes it's hard to tell if it's a real person or AI. If it were a real person, you could vent a little, but with AI, it just pauses and continues."
Despite this, Liu emphasized that the technology itself is neutral. The root of the most criticized spam calls lies not in AI but in information leaks. "For example, knowing precisely when a user has just bought a house and needs renovation or just had a baby and needs formula—the problem still lies in data breaches. AI shouldn't take the blame. It only makes the calling process cheaper. For ordinary users, the solution is to register on fewer websites and apps to minimize privacy risks."
In his view, AI objectively helps companies save costs and improve efficiency, making it an inevitable trend for the future, and investors are generally optimistic about this sector. "Statistically, 90% of consumer calls involve very simple queries, which AI customer service can handle effortlessly. The widespread adoption of AI is unstoppable."
Chen Yan shared similar observations, noting that many companies have rushed into developing AI customer service this year. Some products are still incomplete, with only a basic framework, yet they are already being marketed. While all are labeled as AI, their intelligence levels vary—some can learn continuously, while others fail to recognize slight variations in wording. Additionally, some clients purchasing these products aren't fully aware of the differences, leading to mixed consumer experiences with AI customer service.
Liu Xingliang believes that in the unstoppable wave of AI, companies must balance marketing efforts to avoid alienating consumers and mitigate risks of data leaks and misuse. At the same time, they should develop smarter products to ensure AI customer service is truly effective and not "clueless." "Previously, AI relied heavily on human trainers. With the emergence of large models and ChatGPT, the industry is thriving and will continue evolving toward greater convenience and intelligence."