11-19 更新 实时
首页/ 科技公司/ 正文

企业应把 AI 押注放在哪里

科技公司 来源:The Information AI 2026-11-13 22:41 0 阅读 机器翻译

在推动全公司范围采用 AI 数年之后,企业如今正面对由此产生的成本。随着董事会层面的讨论转向投资回报率,企业领导者正在更严格地审视该为哪些 AI 投入提供资金并扩大规模。

为了解高管在 AI 投资前应该提出哪些问题,The Information 的金融编辑 Meredith Mazzilli 与三位从不同视角看待 AI 的专家进行了交流。

重新设定优先级

在 Meta Platforms,Mark Zuckerberg 采取了实验优先的做法,将采用 AI 列为全公司的优先事项。

“他基本上说的是,‘今年将是所有人都采用 AI 的一年。’我们告诉员工,‘找出什么行得通,去尝试、去失败,’”Heidi Rhodes 回忆道。

这种实验自由帮助 Meta 识别出哪些工作流能从 AI 中获益,并围绕这些工作流实现标准化。起初,Token 消耗并不是问题;Rhodes 承认,并非每个组织都负担得起这种奢侈。

“2023 年时是这样:不管你遇到什么问题,都抹点 AI 上去。然后我们收到了 Token 账单,”她说。“现在各组织在说,‘除非你能向我展示真实的 ROI,否则我们就到此为止。’”

其他公司则采取了更为谨慎的路线。在监管严格的金融服务业运营的 Jack Henry & Associates 从治理入手,建立了防护机制,用于评估和批准 AI 工具,并确定员工可以访问哪些数据,该公司 CFO Mimi Carsley 表示。

在 PwC,挑战与其说是缺少想法,不如说是判断哪些想法值得推进。PwC 各部门的员工都在拥抱 AI,探索将其融入工作的新方式。Karen Young 表示,决定是否为某个想法投入资金的因素仍然是客户,与 AI 出现之前并无不同。

她说:“我们考虑的是客户在寻找什么样的解决方案,以及我们如何[将 AI 工具]规模化,为他们创造价值。”

控制支出

Carsley 表示,在 Jack Henry,为全部 7,300 名员工提供 Microsoft Copilot 许可可能增加 300 万美元成本。因此,公司会定期重新评估谁需要访问权限。

她说:“如果有人没有在使用自己的许可,我们会去和他们谈,指导他们用起来。如果他们仍然不用,我们实际上会把它收回,并在组织内重新分配。”

Young 表示,PwC 也已转向培训员工,让他们更高效地使用这些工具。

她说:“我们引导他们出于正确的理由,为正确的模型选用正确的工具。”

衡量真正重要的事

组织为 AI 支出排定优先级的一种方式,是不只盯着效率提升,还要看这些提升如何被加以利用。举例来说,Young 提到一家客户改造了 14 个财务流程,将预测和情景规划从数周缩短到几分钟。

“真正的问题是:财务部门的那些员工现在如何更战略性地利用自己的时间?”她问道。

Carsley 表示,企业还需要考虑如何利用这些节省下来的时间,以及能否将投资重新部署到其他领域。

“如果有人用 AI 起草一封邮件,省下了五分钟,那你是去拿了杯咖啡,还是用这五分钟做了点什么?”她说。“如果你没有把省下的时间重新投到另一项生产力或效率提升上,那这笔支出真的让你获得了什么吗?”

与此同时,Rhodes 提醒不要将效率提升当作裁减人力的借口。她提到一家客户,在 AI 接手了数小时的研究工作后,其销售团队的效率提升了 20% 到 30%。

“你是要裁减人员?还是既然你拥有了世界上最高效的销售人员之一,就出去招更多人?”她问道,并补充说她倾向于后者。

不过归根结底,AI 优先级的排序取决于一家公司独有的优势。

“问问自己,你在哪方面有优势?然后把 AI 作为加速器应用到那里,”Carsley 建议道。“做你作为一个组织最擅长的事,然后把它提升 10% 到 20%。”

译文由英文原文机器翻译,术语按行业惯例统一,可在文末展开原文核对。

查看英文原文(Where companies should place their AI bets)

After several years of pushing organizationwide AI adoption, companies are now confronting the cost. As boardroom conversations shift to return on investment, business leaders are taking a harder look at which AI bets to fund and scale.

To understand what questions executives should be asking ahead of AI investments, Meredith Mazzilli, The Information’s finance editor, spoke with three experts approaching AI from different perspectives.

  • Karen Young, chair of the U.S. board of partners and U.S. sales leader, PwC
  • Mimi Carsley, chief financial officer and treasurer, Jack Henry & Associates
  • Heidi Rhodes, global director, enterprise GTM customer success and sales enablement, Meta Platforms

Resetting Priorities

At Meta Platforms, Mark Zuckerberg has taken an experimentation-first approach, making AI adoption a priority across the company.

“He basically said, ‘This is going to be the year that everyone is going to adopt AI.’ We told our employees, ‘Figure out what works and go fail,’” recalled Heidi Rhodes.

That freedom to experiment helped Meta identify which workflows benefited from AI and standardize around them. Initially, token consumption wasn’t a concern, a luxury Rhodes acknowledged not every organization can afford.

“In 2023, it was: Any single problem that you have, rub some AI on it. Then we got our token bills,” she said. “Now organizations are saying, ‘Unless you can show me real ROI, we can end the conversation there.’”

Other companies took a more cautious route. Jack Henry & Associates, which operates in the highly regulated financial services industry, started with governance, establishing guardrails for assessing and approving AI tools and determining what data employees could access, said CFO Mimi Carsley.

At PwC, the challenge is less a shortage of ideas than deciding which ones are worth pursuing. Employees across PwC are embracing AI and exploring new ways to integrate it into their work. Karen Young said the deciding factor in funding ideas remains the client, much as it did before AI.

“We think about the solutions our clients are looking for and how we can scale [our AI tools] for their benefit,” she said.

Controlling Spend

At Jack Henry, providing Microsoft Copilot licenses to all 7,300 associates could add $3 million in costs, said Carsley. The company therefore regularly reassesses who needs access.

“If someone’s not using their license, we’ll go talk to them and coach them up. If they’re still not using it, we actually seize it back and redeploy it elsewhere in the organization,” she said.

Young said PwC has also shifted to training employees to be more efficient with their usage.

“We lead them to use the right tool for the right model for the right reason,” she said.

Measuring What Matters

One way organizations can prioritize AI spending is to look beyond efficiency gains to how those gains are put to use. By way of example, Young cited a client that transformed 14 finance processes, cutting forecasting and scenario planning from weeks to minutes.

“The real question is: What are those employees in finance now doing to be more strategic with their time?” she asked.

Carsley said companies also need to consider how they are using those time savings and whether they can redeploy investment elsewhere.

“If someone’s gonna use AI to draft an email and you save five minutes, like, did you go get a cup of coffee or did you do something with the five minutes?” she said. “If you’re not redeploying that savings to another productivity or efficiency, did you really gain anything by the spend?”

Rhodes, meanwhile, cautioned against treating efficiency gains as an excuse to reduce headcount. She cited a client whose sales team became 20% to 30% more productive after AI took over hours of research.

“Do you make a reduction in humans? Or now that you have one of the world’s most efficient salespeople, do you go out and hire more?” she asked, adding that her vote is for the latter.

Ultimately, though, AI prioritization comes down to a company’s unique advantage.

“Ask yourself, what do you have an advantage in? Then apply AI to that as an accelerant,” advised Carsley. “Do the thing you do best as an organization, then turn it up 10% to 20%.”

正文保留英文原文,标题与摘要由机器翻译。

本文由采集器自动同步

原文发布于 The Information AI(Where companies should place their AI bets),标题、摘要与正文由机器翻译,术语按行业惯例统一,未经人工逐句校订。

阅读原文 ↗