纳丹·尼勒卡尼:Infosys 的 AI 转型与企业传承 Norges Bank Investment Management 2026-05-15

创始人身份与传承挑战

纳丹·尼勒卡尼: 谢谢你,Shilpi,能来到这里真的很棒。我也要感谢 Nikolai 给我的这次机会。我想和大家分享一点我的故事,以及它与今天主题的联系。我是 Infosys 的创始人之一,这家公司成立于 45 年前。在这 45 年里,除了三年之外,公司里一直都有创始人在场。而那三年没有创始人的时间里,公司经历了一段动荡的时期。2017 年,因为当时出现了一些情况,我不得不回去——我当时已经离开了,我在 UIDAIAadhaar 做了所有工作。我本以为回去几个月就好,结果九年过去了,我还在那里。所以我就是最后的创始人,你知道的,就像最后的恐龙,最后的莫西干人,最后的创始人。

我的脑海里一直在纠结一个问题:我该如何过渡到一个非创始人的领导者?换句话说,虽然我们有正常的继任计划(即用一个 CEO 替换另一个 CEO),但当涉及到用非创始人取代创始人时,其复杂程度要高出一个数量级,因为它关乎到价值观、文化等等。所以我已经在纠结如何完成这次过渡。

Original English

Nandan Nilekani: Uh thank you Shilpi and it's really wonderful to be here and I want to thank Nikolai for this opportunity. I thought I'll tell you a bit of my story and the connection to today's topic. I am one of the founders of Infosys, a company set up 45 years back and in these 45 years except for three years there has been one or more founder in the company and the three years where there was no founder the company went through a tumultuous time. So in 2017 because of a certain situation that had arisen I had left by then I had done all the work at UIDAI Aadhaar and all that I had to go back I thought I was going back for a few months 9 years later I'm still there so I am the last founder you know like the last dinosaur the last moan the last founder. So I was already in my mind grappling with the issue, how do I do a transition to a leader who will not be a founder? In other words, while we have normal succession planning where we replace one CEO with another, when it comes to replacing a founder with a non-founder, it's a order of magnitude more complicated because it's all about, you know, the values that you have and and and the culture and so on. So I was already grappling with this question of how do I do this transition and then this whole AI thing came along.

AI 转型带来的 existential 危机与机遇

纳丹·尼勒卡尼: 然后,整个 AI 热潮随之而来,人们开始质疑我们存在的意义。如果软件开发要在云端、或者用 CursorCodeex 或其他工具来完成,那为什么还需要我们这些人?你们还有存在的必要吗?我们突然面临这个存在主义的问题。但我意识到,作为一个乐观主义者,我总是寻找积极的一面,这可能是一个机遇。我说,也许在我作为主席向未来领导者过渡的这个节点,我可以利用这个机会再次将这些价值观植入公司,让它能持续更久。

在某种意义上,我让 Infosys 的 AI 转型也变成了一种改变文化的方式,以创造一种紧迫感并提升速度。这是一家拥有 30 万员工、年营收 200 亿美元、业务遍及 40 多个国家的公司,因此在这种公司里提升速度是非常困难的。我的第一份工作就是如何实现这一点。我还意识到,我们必须通过确保在世界范围内拥有战略自主权来做到这一点。

Original English

Nandan Nilekani: And then this whole AI thing came along and people started questioning why we exist. If software development is going to be done on cloud or cursor or codeex or whatever then why do we even need you people? Do you need to exist? So we had suddenly come to this existential question. So then I realized you know I also then realized that maybe this was an opportunity for me because being an optimist I always look for the upside. I said maybe this point where there will be a transition to a future leader at some point as the chairman. I could also use this opportunity to embed the values once again into the company so that it would be instead for a long time. And so in some sense I made the AI transition of Infosys actually become also a way to change the culture to create a sense of urgency to get speed and getting speed this is a company with 300,000 people and 20 billion in revenue so to get speed in this company is you know over 40 countries so I think for my first job was how do we get this and then I also realized that we have to do do this only by ensuring that we have strategic autonomy in the world.

AI 时代的定位:编排与场景化

纳丹·尼勒卡尼: 如何在 AI 世界中保持相关性?有几件事对我们来说很突出。第一,像 Infosys 这样的公司,其作用应该是编排企业内部的所有事物,因为企业需要对它使用的所有工具进行编排。这还关乎帮助我们的客户在选择 AI 工具时保留可选性,因为我坚信 AI 领域的领导者会不断更迭。会有新的公司、新的做事方式出现,这要求企业必须灵活敏捷,能够适应这些变化。因此,我们能提供的关键价值主张之一,就是如何在技术选择上实现可选性。

第三件事是,我们意识到企业上下文(Enterprise Context)非常重要。企业是由其上下文、规则、政策、流程、数据定义的,其中一些可能是员工头脑中的隐性知识,或者是贴在电脑上的便签。所以,我们意识到,整合企业上下文对我们的客户非常有价值。我们摸索出了在这个新世界中,我们的价值主张和相关性应该是什么。

Original English

Nandan Nilekani: How to be relevant in the world of AI and a few things that stood out to us. One was that the role of a firm like Infosys would be to orchestrate all the things in a company in an enterprise because a enterprise needs orchestration of all the various tools it uses and so on. It was also about helping our clients retain optionality in the choice of the AI tools because I firmly believe that the leaders in AI will keep changing. There'll be new companies, new new ways of doing things. All that is going to come in and that's going to be requires therefore companies to be nimble and agile in being able to adjust to these changes. So one of the key value ads we could offer was how do we how do we allow optionality in the choice of our technologies. And the third thing is we realized that enterprise context is so important that an enterprise is defined by its context, its rules, its policies, its its processes, its data and some of it would be tacet knowledge in people's heads, post-it notes stuck on the computer. So fundamentally getting enterprise context together we realized was something very useful for our customers. So we figured out how to think about what would be our value proposition and relevance in this new world.

组织对齐:步调一致的重要性

纳丹·尼勒卡尼: 同样重要的是,我们利用这个机会来提升速度,因为我们需要非常迅速地行动。我们必须推动整个公司——30 万人,培训每个人掌握这些新技能,学习如何使用代理(Agent),学习如何将代理与人类结合起来做事。这需要大量的工作。这也是打破公司孤岛的一个机会,因为之前有不同的孤岛,人们做着不同的事情。我们说,让我们把 AI 转型作为一种打破孤岛、让公司每个人都以同样方式工作的方式。这花了一些精力,因为人们对自己做事的方式相当执着。

但我坚信,除非公司步调一致,用一个声音说话,否则你无法获得 AI 的好处和生产力。每个人都必须朝着相同的目标努力。AI 不仅仅是应用人工智能,它是个体对齐(Aligned Individuals)。公司里的所有个人必须朝着同一个方向前进,否则 AI 可能会让事情变得更糟。一个运作失调的公司如果用上 AI,只会加速走向毁灭。如果你给我发一封邮件,我用 AI 把原本一段话扩充成十段发给你,你再用 AI 把这十段缩减成一段。我们得到了什么?我们双方都用了 AI,但公司的生产力并没有提高。如果你有一家公司,里面有不同的派系,那么你所知道的,一个业务部门的代理群就会和另一个业务部门的代理群发生争斗,而且是 24/7 全天候的。

因此,我们必须非常小心,不要以一种运作失调的方式实施 AI,因为那可能是适得其反的。所以我学到的教训是:利用这次 AI 转型,让公司步调一致,确保创新从公司的各个部分捕获到一个共同的运作方式——一个共同的织物(Fabric)中。我现在感到自信,虽然未来某个时候会有向非创始人的过渡,但 AI 实际上给了我们一个重新校准、重新激发活力、减少官僚主义、打破孤岛并带领员工迈向新高度的机会。我很有信心,无论何时我们进行 AI 转型,我们也会完成主席和创始人的传承。非常感谢。

Original English

Nandan Nilekani: But equally importantly I think we use this as an opportunity to get speed because we had to move very quickly. We had to move the whole whole company 300,000 people. We had to train everyone to be ready with these new skills. We had to learn how to do agents. We had to learn how to combine agents and humans in doing things. All this required a lot of work. It was also an opportunity to break down the silos of the company because there were different silos. People are doing different things. And we said let's use the AI transformation as a way to completely break down the silos and have everyone in the company working on the same have the same way of doing things. And that took some effort because people were quite possessive about the way they did things. But we have used used this to actually break that down because I firmly believe that you cannot have AI benefits and AI productivity unless the company works with one voice with in unison. Everybody has to work to the same goals. And for I think AI is not just applied artificial intelligence. It's aligned individuals. All the individuals in the company have to be going in the same otherwise AI could make it worse. A dysfunctional company can go into warp speed with AI. You know if I'm sending a mail to somebody I may I may write a one paragraph email use AI to make it 10 paragraphs send it to that guy and he use AI to reduce 10 paragraphs to one paragraph. What have we achieved? Both of us have used AI on either side but the company's productivity has not gone up. If you have a company which is having camps and different people, then for all you know, agent swarms of one business will be fighting with the agent swarms of another business and it'll be doing it 24 by7. So in other words, we have to be very careful that we don't implement AI in a way that is dysfunctional because that could actually be counterproductive. And therefore I think the lesson I learned is use this AI transformation to also get the company to march to a single drummer to get everyone aligned to do the things that they they had to do and to make sure that innovation was captured from every part of the company onto a common way of doing things a common fabric. So I feel now that while there will be at some time in the future a transition to a non-founder the AI has actually given us an opportunity to also recalibrate and re-energize and reduce bureaucracy reduce silos and take people to a different level. So finally I think I feel confident that whenever we do the AI transition we will also do a chairman and founder transition. Thank you very much.

📌 文中提及的人物和组织

人物: Nandan Nilekani

公司/组织: Infosys, UIDAI

关键字: ai-transformation organizational-leadership enterprise-context corporate-culture ai-orchestration