背景介绍：
随着科技的飞速发展，客户服务行业正经历着前所未有的变革。家庭音响系统制造商 Sonos 在面对产品故障引发的客户愤怒时，引入了由 OpenAI 执行董事长布雷特·泰勒共同创立的 Sierra 公司提供的生成式 AI 客服机器人，这一创新举措展现了 AI 在客服领域的潜力。然而，客服行业长期因过度依赖科技导致顾客体验不佳而饱受诟病。如今，随着生成式 AI 技术的不断进步，初创公司和科技巨头纷纷涌入这一领域，试图通过 AI 客服助手改善客户体验，并有望彻底改变客户服务业的面貌。
Can artificial
intelligence rescue customer service?
AI
能拯救客服行业吗？
The adoption
of AI is surging in call centres
越来越多客服中心开始拥抱人工智能
It’s not easy being a
customer-service agent
—particularly when those customers are so angry with a
product that they want to yell at you down the phone. That’s the sort of rage
that Sonos, a maker of home-audio systems, encountered in May when it released
an app update so full of
glitches
it caused its share price to plunge.
做客服可不是件轻松的事——尤其是当顾客对产品极为不满，恨不得隔着电话对你大喊大叫的时候。今年5月，家庭音响系统制造商 Sonos 就遭遇了这样的愤怒情境。当时，他们发布的一款应用更新版本存在诸多漏洞，直接导致其股价暴跌。
One of the agents dealing with the ensuing customer
tirades was a rookie. But not a human one. Prior to the debacle, Sonos had
hired Sierra, a startup co-founded by Bret Taylor, the chairman of OpenAI, to
provide it with a customer-service bot powered by generative
artificial intelligence
(AI).
在应对随后客户一连串责难的过程中，有一位客服是初出茅庐的新手，但这位客服并非人类。早在这场事故发生之前，Sonos 就已聘请了由 OpenAI 执行董事长布雷特·泰勒共同创立的初创公司 Sierra 为其提供由生成人工智能（AI）驱动的客户服务机器人。
It could have been a disaster; the only thing worse than a malfunctioning product is being trapped in an automation prison by a robot giving you the runaround. Yet the bot beat expectations. After digesting Sonos’s technical materials, it came up with its own workaround for one of the problems with the Sonos app.
这可能原本被视为一场潜在的灾难；因为没有什么比产品出现故障更令消费者沮丧的了，而比这更糟糕的是被一个机器人无休止地困在自动化回复的循环中。然而，这次，这个机器人却展现出了超乎预期的能力。在深入学习了 Sonos 的技术材料后，它针对 Sonos 应用程序的一个具体问题，独立提出了有效的解决方案。
Customer service is one of the few industries where the
use of generative AI is already taking root. In a survey of customer-service
executives published earlier this year by Gartner, a research firm, almost half
said that AI customer assistants would have a significant impact on their
organisations in the next 12-18 months.
客户服务是生成式 AI 技术应用已初步显现成效的少数行业之一。研究公司 Gartner 今年早些时候发布的一项针对客户服务高管的调查显示，近半数的受访者认为，在未来12至18个月内，AI 客服助手将对他们的公司产生重大影响。
Startups and established tech firms alike have launched a volley of new products at the industry that promise to transform customer service—and millions of jobs.
初创公司和成熟的科技公司纷纷在该行业推出了一系列新产品，这些创新有望彻底改变客户服务业的面貌，并影响到数百万个工作岗位。
Customer service is a big industry. Most companies have
some sort of customer support, whether in-house or outsourced to call centres.
In America alone there are almost 3m customer-service workers, according to the
Bureau of Labour Statistics.
客户服务是一个庞大的行业。大多数公司都设有某种形式的客户支持部门，无论是内部自建还是外包给专业的呼叫中心。据美国劳工统计局数据显示，仅在美国，客服人员的数量就近300万。
At a median salary of around $40,000 a year, that works out at roughly $120bn in wage costs. Many more work in call centres in places like India and the Philippines, where these jobs are seen as ladders to the middle class.
按年薪中位数约4万美元计算，这一行业的工资成本高达约1200亿美元。而在印度和菲律宾等地的呼叫中心，客服人员的数量更为庞大，这些工作被视为许多人迈入中产家庭的阶梯。
In recent years, however, the industry has become
notorious
for driving customers mad with its use of technology. Its poor
reputation is deserved, reckons Andy Lee, co-founder of Alorica, an American
contact-centre business with 100,000 employees.
然而，近年来，客服行业因过度依赖科技而声名狼藉，常常让顾客感到抓狂。拥有10万名员工的美国联络中心企业 Alorica 的联合创始人安迪·李认为，这一行业的恶名并非空穴来风。
It is expensive for firms to use humans to solve their customers’ problems, so they make the process as
cumbersome
as possible by forcing them to press a bewildering combination of numbers or chat with a bot that regurgitates generic responses—a strategy known as “
deflection
”.
对于公司而言，使用人类来解决客户问题成本高昂，因此，这些公司往往会通过设计复杂的流程来减少人工介入，比如迫使客户按下一系列令人困惑的数字组合，或是与只会重复固定答案的机器人交流，这种做法被称为“注意力转移”策略。
Once human agents are involved, it is in the financial interests of outsourcing firms to make the process as labour intensive as possible, raising costs and frustrating everybody.
一旦涉及人工代理，外包公司出于经济利益考虑，会使流程尽可能劳动密集，这不仅提高了成本，也让双方都感到沮丧。
Now entrepreneurs and investors are betting that
generative AI can make things less awful. Funding for startups developing
customer-service tools that use generative AI reached $171m globally in the
third quarter, up from $45m in the same period last year, according to
PitchBook, a data gatherer (see chart).
如今，企业家和投资者正寄望于生成式 AI 能够改善这一状况。根据数据收集公司 PitchBook 的数据，第三季度，全球开发使用生成式AI客服工具的初创公司融资额达到了1.71亿美元，这一数字远高于去年同期的4500万美元。
This month Crescendo, co-founded by Alorica’s Mr Lee, raised money at a valuation of $500m. Sierra, which raised funds at a valuation of about $1bn in January, is now said to be seeking more at a valuation of $4bn, raising eyebrows even among some AI-crazed venture capitalists.
本月，由 Alorica 的安迪·李联合创立的 Crescendo 公司以5亿美元的估值成功完成融资。据悉，今年1月以约10亿美元估值完成融资的 Sierra 公司，目前正在寻求以40亿美元的估值进行更多融资，这一举动甚至让一些对人工智能充满热情的风险投资家也感到惊讶。
It is not just startups entering the field. Tech titans
like Alphabet, Amazon and Microsoft are bringing generative AI to their customer-service offerings. Software firms
like Salesforce are, too. Some companies are also using large language models
like those produced by OpenAI to create their own customer-service bots.
进入这一领域的不仅仅是初创公司。Alphabet、亚马逊和微软等科技巨头正积极将生成性人工智能融入其客户服务产品中。Salesforce 等软件公司同样如此。一些公司还利用 OpenAI 制作的大型语言模型来打造自己的客户服务机器人。
Unlike their rote-learning predecessors, generative-AI
bots do not
regurgitate
canned answers to narrow questions. Instead, they
create their own responses informed by the firms’ training materials and
previous customer-service interactions.
与那些只能死记硬背的人工前辈不同，生成式人工智能机器人不会局限于提供狭隘问题的现成答案。相反，它们会根据公司的培训材料和以往的客户服务互动，自主创造个性化的回应。
