The golden age of unbridled spending on AI software might be behind us, as vendors say it’s a lot harder to make a sale than it used to be.随着供应商表示如今销售难度远超以往，人工智能软件无节制投入的黄金时代或许已成过去。Last year represented something of a boon era for vendors peddling AI apps. Spurred by board-level mandates, corporate FOMO and an aggressive campaign from tech giants about the world-changing capabilities of AI agents, enterprises were spending willingly and wildly—an estimated total of more than $1.249 trillion in software, according to research and advisory firm Gartner.去年对于推销人工智能应用程序的供应商而言，可谓繁荣时期。 在董事会层面的指令推动下，受企业“错失恐惧症”（FOMO）心理以及科技巨头关于人工智能代理将改变世界的激进宣传影响，企业纷纷慷慨解囊、大肆采购。 据研究和咨询公司高德纳（Gartner）估计， 去年企业在软件方面的支出总额超过1.249万亿美元。Alex Levin, co-founder and chief executive of AI-powered customer service startup Regal, said until recently he could make an enterprise sale with a single demo—a phenomenon he called “shocking,” given how big companies can sometimes take one to two years to make a purchase. Last year, Levin said he was closing deals in as little as 60 or 90 days.人工智能驱动的客户服务初创公司Regal的联合创始人兼首席执行官亚历克斯·莱文表示， 直到最近，他只需一次演示就能达成企业销售交易——他称这一现象“令人震惊”， 因为大公司有时需要一到两年才能完成采购。 去年，莱文称自己最短能在60天或90天内就敲定交易。That isn’t the case anymore.但如今情况已非如此。Vendors say big companies have become more cautious about what they buy. They’re taking longer to evaluate solutions, involving more internal stakeholders from legal and finance teams, and placing more emphasis on the kind of financial returns they might get out of the investment. The breakneck pace of AI innovation—like recent updates around Anthropic’s Claude—is also making potential customers wary of sales commitments.供应商表示，大公司如今在采购方面变得更加谨慎。 它们评估解决方案的时间更长，会让法律和财务等更多内部利益相关者参与进来， 并且更加注重投资可能带来的财务回报。 人工智能领域创新速度惊人——比如Anthropic的Claude近期推出的更新——这也让潜在客户对销售承诺心存疑虑。“There was a period where the early adopters were moving very fast on really interesting technology and that piece has slowed down,” Levin said. The typical time for completing a sale is now about six months, he said.“早期采用者曾对非常有趣的技术快速行动，而如今这一阶段已经放缓，”莱文说。 他表示，现在完成一笔销售交易通常需要约六个月时间。“Everyone is a bit more cautious,” said Craig Roth, a vice president analyst at Gartner. “I think reality has set in.”“大家都更加谨慎了，”高德纳副总裁分析师克雷格·罗斯说， “我认为现实已经显现。”Early adopters who rushed into AI pilots and even deployments last year often hit a wall and learned some hard lessons. It wasn’t necessarily because the technology didn’t work, but because they found they didn’t have the right guardrails or didn’t fully understand the reality of the business process they were trying to automate, Roth said. Critically, they found it was hard to measure financial returns. And what they could measure wasn’t overwhelmingly impressive, he said.去年匆忙开展人工智能试点甚至部署的早期采用者，常常遭遇困境，吸取了惨痛教训。 罗斯说，这未必是因为技术不起作用， 而是因为他们没有建立恰当的管控机制，或者没有充分理解他们试图实现自动化的业务流程的现实情况。 至关重要的是，他们发现很难衡量财务回报。 而且，他们能够衡量的回报也并不十分可观，他说。In a Gartner survey released in April 2025, only 11% of customer service and support leaders said generative AI met their primary business objective—striking, since customer service has emerged as one of the most mature areas for deploying the technology.在高德纳2025年4月发布的一项调查中， 只有11%的客户服务与支持负责人表示生成式人工智能达到了他们的主要业务目标——这一结果令人震惊， 因为客户服务已成为部署该技术最为成熟的领域之一。Businesses are certainly continuing to invest in AI tools. Gartner expects software spending to grow 14.7% this year to about $1.434 trillion. But now that businesses are mature enough to understand those potential roadblocks, they are taking longer to evaluate and being more critical of potential solutions, Roth said.企业当然仍在继续投资人工智能工具。 高德纳预计，今年软件支出将增长14.7%，达到约1.434万亿美元。 但罗斯表示，如今企业已经足够成熟，能够理解这些潜在障碍， 因此会花更长时间进行评估，并对潜在解决方案更加挑剔。“Overall, we’ve gotten a lot more disciplined in making sure that we understand what the outcome of a specific purchase is, not just following the hype,” said Kyle Chu, senior manager of Business Intelligence at phone-accessory maker PopSockets.“总体而言，我们在确保理解特定采购的成果方面变得更加自律， 而不再仅仅是跟风炒作，”手机配件制造商PopSockets的业务智能高级经理凯尔·朱说。Kathy Kay, chief information officer and executive vice president of global financial services firm Principal Financial Group said she’s taking her time to think deeply about whether a given vendor will even be around or be useful in a few years, “which could make it seem like the sales cycle is longer to a company calling on us,” she said.全球金融服务公司信安金融集团的首席信息官兼执行副总裁凯西·凯表示，她会花时间深入思考某个供应商在未来几年是否还会存在，或者是否还有用， “这可能会让给我们打电话的公司觉得销售周期变长了，”她说。Big tech players like Microsoft and Google, who have increasingly moved to bundle generative AI tools into existing broader offerings, are less affected. But, for smaller companies whose AI offerings are their raison d’être, there are pros and cons of the new era.像微软和谷歌这样的大型科技企业，它们越来越多地将生成式人工智能工具捆绑到现有的更广泛产品中， 受到的影响较小。 但对于那些以人工智能产品为核心业务的小公司而言，这个新时代有利有弊。It might be harder to get in the door, but once they do, they find enterprises want to use the solution more broadly, and get everything they can out of it.进入企业大门或许更难了，但一旦成功进入，它们发现企业希望更广泛地使用解决方案， 并从中获取最大价值。Many say the difference manifests in terms of who shows up in the procurement meetings.许多人表示，这种差异体现在采购会议的参会人员上。Are Traasdahl, founder and CEO at AI-powered retail data platform Crisp, said he’s engaging more with non-tech C-level executives than ever before, and they want to ensure the solution works consistently across the business, rather than just speaking with a leader in one siloed part of the company.人工智能驱动的零售数据平台Crisp的创始人兼首席执行官阿雷·特拉斯达尔表示，他比以往任何时候都更多地与非技术领域的高管进行接触， 而且这些高管希望确保解决方案能在整个企业内稳定运行， 而不仅仅与公司某个孤立部门的负责人交流。Oisin Hanrahan, co-founder and CEO of AI-powered manufacturing software startup Keychain, agreed.人工智能驱动的制造业软件初创公司Keychain的联合创始人兼首席执行官奥辛·汉拉汉对此表示认同。“In 2025, business unit leaders could make decisions,” he said. “Now, if you go into the conversation, you see someone from finance show up. You see someone from legal show up in addition to the business unit, which tells you ‘Hey, we’re taking this much more seriously.'”“2025年，业务部门负责人可以自行做决策，”他说， “如今，如果你参与对话，会发现财务部门的人来了。 法律部门的人也来了，此外还有业务部门的人，这表明‘嘿，我们对这件事更加重视了’。”Hanrahan added, “The threshold for a pilot has gone up across the board.”汉拉汉补充道：“整体而言，开展试点的门槛提高了。”长按上方二维码，下载英语日报APP
夜雨聆风
