Workday的最后一个工作日？人工智能与企业软件的未来

没人喜欢在Workday里做这些乱七八糟的任务。

No one likes doing, like, these crazy tasks inside of Workday.

没哪个员工愿意跟Workday打交道。

No employee likes interacting with Workday.

谁都不想登录这个门户。

No one wants to go into this portal.

如果你能出去解决这些问题，那可是个巨大的市场机会。

And so if you can go out there and solve these problems, there's a huge pie to go after.

想想你得去哪儿拉这些客户——要是人家已经用了Workday、ServiceNow或者Salesforce，想把他们挖走简直难如登天。

If you think about where you have to go capture these customers, if someone has already become a customer of Workday or already become a customer of ServiceNow or Salesforce, it is so hard to rip them off.

所以这些毛收入留存率才这么高。

That's why these gross dollar retentions are what they are.

它们之所以是行业标杆，是有道理的。

Like, they're the gold standards for a reason.

我们正看到全球最护城河深的公司也开始出现裂痕，这真是个激动人心的时刻，比赛已经开始了。

We're seeing the, like, cracks in the most defensible businesses in the world, and it's a really exciting time, and and the race is on.

要取代企业软件里这么根深蒂固的系统，得需要什么条件？

What does it take to replace one of the most entrenched systems in enterprise software?

过去二十年，像Workday这样的平台已经成了企业管理员工、数据和运营的基石。

For the past two decades, platforms like Workday have become the backbone of how companies manage people, data, and operations.

它们早就深深扎进去了，护城河特别厚，想拔都拔不动。

They're deeply embedded, highly defensible, and nearly impossible to rip out.

但平台的更替会彻底改写规则。

But platform shifts change the rules.

就像2000年代云计算彻底改变了企业软件一样，现在AI又带来了新机会，让我们重新思考：这些系统到底该怎么运作，甚至该不该还以现在的形态存在。

Just as cloud transformed enterprise software in the 2000s, AI is creating a new opportunity to rethink how these systems work and whether they should exist in their current form at all.

现在的问题不再是这些系统有没有价值，而是它们能不能跟得上变化的速度。

The question now is not whether these systems are valuable, but whether they can evolve fast enough.

埃琳娜·伯格采访了a16z企业团队的合伙人乔·施密特。

Elena Burger speaks with Joe Schmidt, partner on the enterprise team at a16z.

大家好，欢迎收听《观察局势》。

Hello, everyone, and welcome to monitoring the situation.

我这儿跟着的是第四代乔·施密特。

I am here with Joe Schmidt the fourth.

乔是

Joe is a

谢谢你把这点说清楚。

Thank you for making that delineation.

我很感谢。

I appreciate it.

乔是a16z企业团队的合伙人。

Joe is a partner on the enterprise team at a16z.

我们今天其实想先回答一个关键问题：为什么企业团队最近写了这么多悼文、这么多追思文章？

And we are here to actually, I think, first just answer a big important question, which is why has enterprise team been writing so many obituaries lately, so many eulogies lately?

这些悼文又是悼念谁的呢？

And what have these eulogies been for?

是的。

Yeah.

不是。

No.

谢谢你邀请我来。

Thanks for having me on.

很高兴能来。

Excited to be here.

对。

Yeah.

说这些是悼词挺有意思的，我觉得那篇文章本意是想更平衡一些，可能大家理解得有点偏了。

It's funny calling them eulogies, and I think that this piece was intended to be much more balanced than maybe people have taken it to be.

但如果你想想现在企业软件所处的周期阶段，就会发现，今天那些最大的企业，其实都是上一轮技术变革——也就是从本地部署转向云——的时候建立起来的。

But I think that if you just think about where we are in the cycle right now with enterprise software, a lot of the biggest businesses that are out there today were built in the last technology shift, which was obviously a shift from on prem to cloud.

对大多数面向大企业客户的新公司来说，真正的机会就是抓住这种大的技术转折，围绕它打造自己的平台。

And that's really the big opportunity for most companies that are being built, selling to big enterprise buyers, is basically adopting some big shift and then building their platform around it.

一旦这些核心基础打好了，这些市场就会转向另一种创新，也就是在某个核心后台系统之上，开发各种专门的解决方案。

Once those kind of core primitives have been set, basically, these markets end up moving into a different type of innovation, which ends up being like point solutions that are built on top of some kind of core back end system.

原因在于，如果你仔细想想这些系统到底在做什么，比如Workday，它把公司内部的人力资源流程和一堆业务逻辑都固化在一个关系型数据库里。

And the reason for that is if you think about what these systems are actually doing, and if you take Workday, obviously it codifies your HR business processes and a bunch of business logic internal to your company inside of one relational database.

这里面涉及的东西特别复杂。

And there's a bunch of really complex stuff that happens there.

但员工、HR管理员或者IT人员实际使用时的体验，跟2005年几乎没什么两样。

But the actual experience of that, that an actual employee or an HR administrator or someone in IT experiences, is fundamentally the same today as it was in 2005.

所以如果你跑过去说，‘我要重做一遍Workday’，跟CFO或者CIO这么说，他们肯定会问：‘那它跟Workday以前的版本到底有什么不一样？’

And so if you come in and you say, Hey, I'm going to rebuild Workday, and you tell CHARR that or you tell the CIO that, they're going be like, Okay, but what does it actually do differently than what Workday has done previously?

而过去，你真很难说它能有什么实质性的不同。

And you're going to say, Until now, it would've been very hard to say it was going to do something different.

过去十五年、二十年，几乎所有企业软件公司都在做同一件事：在这些系统上面堆各种小功能模块。

And so what many companies have been doing over the past fifteen, twenty years, and this has been basically all of enterprise software, has been building little point solutions on top of that.

招聘领域有一堆，财务领域也有一堆，诸如此类。

You'll be seeing a bunch of things in recruiting, a bunch of things in finance, and so on and so forth.

久而久之，Workday就成了这些小工具的统一平台。

And Workday has become the home for a lot of those businesses over time.

我们之前聊过他们的一些收购。

Obviously, we talked about some of their acquisitions.

所以现在这个时刻特别有意思，也是为什么我们写了这么多‘悼词’——其实我不太喜欢叫它们悼词，更准确说是机遇分析。对Workday来说是机遇，对初创公司来说也是同样，因为第一次，你真的可以去跟首席人力资源官或者首席信息官说：你们现在用的这个核心系统，完全能被彻底重构，我们能改变你们团队的工作方式，改变员工或客户与这些平台互动的方式。

And so what's really interesting about this moment in time and why we're writing so many obituaries, which I hate calling them that, it's like these are opportunity pieces, and it's an opportunity for Workday just the same as it is for startups, is that for the first time you can actually go to a CHRO or you can go to a CIO and say, The way that this core system works for you today can be so fundamentally different, and we can change the actual way that the work is being done by your team and the way that these platforms are actually interacted with by your employees or your customers.

这才是真正的机会。

That's the opportunity.

我在人力资源领域就看到了这种变化。

I see that in HR.

在IT服务管理领域也一样。

We're seeing that in ITSM.

在客户关系管理领域也是如此。

We're seeing that in CRM.

甚至在那些历来最稳固、最难被撼动的领域里，我们都能看到这种趋势。

We're seeing that in the biggest and most defensible categories, historically defensible categories.

我觉得，一场平台重构的大潮就要来了。

And I think we're going to see a replatforming.

所以你就能明白，为啥今天股市上会是这副景象了——很多投资者都在讨论这事，还特别担心。

And that's why you see, obviously, what's happening in the in the in the stock markets happening today because a lot of investors are talking about that and are worried about that.

很多老牌企业也是这样。

And a lot of and a lot of incumbents are too.

这正是我们写这些文章的原因，这才是真正的机会。

So that's why we're writing these pieces, and that's the opportunity.

对啊。

Yeah.

以前那些SaaS产品，再也别想靠高企的市盈率撑场面了。

So so these SaaS products of yore can't get their, you know, their juicy PE ratios.

它们没法再指望靠一点点增收和夸张的运营杠杆赚钱了。

They can't, you know, bank on incremental revenue and crazy operating leverage anymore.

现在得真刀真枪地干，要么转型，要么拥抱AI，不然就会被那些新兴对手彻底碾压。

They actually have to do real work and, you know, like, pivot and, you know, adopt AI or get outcompeted by, you know, the insurgents who are.

所以我想，咱们干脆打开你的那篇文章，好好聊聊吧。

So with that in mind, I feel like we should pull up your piece and, you know, chat through it.

是的。

Yeah.

那咱们直接切入正题吧。

So let's let's just get into it.

这篇文章叫《Workday的最后时刻》。

And you can this is a it's called Workday's Last Workday.

我们今天早上刚在a16z.news上发了这篇。

We published it on a16z.news this morning.

你开篇就说Workday是企业软件里最重要却最不受待见的产品。

And you opened the piece by calling Workday the most important and least loved product in enterprise software.

我知道你为了这句开场白可没少纠结。

And I know that, you know, I think you agonized over this opening sentence a lot.

能说说你为啥这么写吗？

So can you can you sort of explain the sentence?

这句话到底揭示了企业软件的什么本质？

Like, what does this tell us about the nature of enterprise software?

怎么可能一样东西既至关重要，又完全不招人喜欢呢？

How can something be, you know, crucially important but also, like, unloved completely?

对。

Yeah.

我的意思是，这可能跟之前我说的有关——这些系统是什么时候、怎么建起来的，还有它们到底是为谁设计的。

Mean, I think maybe it's it goes back to what I just talked about in terms of, like, when when and how these were built and maybe who these were built for.

我猜看这个视频的人里，肯定有人日常用过Workday，说不定还有人是Workday的专业用户，比如顾问或者天天跟这系统打交道的HR管理员。

And I I think, you know, I I'm sure some people that are are watching this have interacted with Workday in their in their in their day to day, and maybe some people are Workday professionals, whether it's consultants or HR admins that actually deal with this on a day to day basis.

但Workday刚出来的时候，那会儿都二十年前了，它的架构现在看也早就过时了。

But when Workday was built, this is now a twenty year old business, and the architecture is obviously 20 old.

它当时算是对PeopleSoft、Oracle那些老系统的一种升级，但本质上还是为内部团队设计的。

It was a better version of kind of the PeopleSoft, Oracle, you know, businesses that existed back in the past, and it was built for those internal teams.

对吧？

Right?

所以它在当时确实很好地解决了它要解决的问题。

And so I think it did a good job of solving for the problem that it had at the time.

我可不会坐在这儿说Workday一无是处。

And I'm not gonna sit here and say there's nothing good about Workday.

它的客户留存率高达97%。

It has 97% gross dollar retention.

它在这个领域就是标杆。

It is the name in the category.

但我真觉得，员工用的那个界面，跟现在顶尖软件的样子差远了。

But I really don't think that the user interface for employees is anything close to what great software looks like today.

对吧？

Right?

我们Intrace Horowitz这儿用的就是Workday。

Like, I am we we are a Workday shop here at Intrace Horowitz.

我本人也登录过Workday。

Like, I've logged into Workday.

真的不行。

No.

真的没有。

No.

我跟你说真的。

I'm not kidding you.

过去十二个月里，我碰了三次壁。

Three times in the past twelve months.

我是Intrace和Horowitz这边负责保险的人。

I am the insurance person here at Intrace and Horowitz.

我根本不知道自己有哪些福利。

I have no idea what my benefits are.

完全不知道。

No idea.

其实没多久前，我专门去Workday里想查一下我的薪酬信息。

I went in actually not too long ago to actually try to find my compensation information in Workday.

花了六分半钟。

Took me six and a half minutes.

我是个科技投资人。

I am a technology investor.

所以啊，这系统根本不是为这个设计的。

So yeah, it's just not built for that.

我跟越多的人聊，就越发现这体验彻底烂透了，而机会就在于重新做一遍。

And I think that the more I talk to people, the more I realized that that experience is broken and the opportunity here is to build that again.

我觉得这特别重要，因为里头的信息太关键了。

So I think it's really important because it has very important information.

我说的那些东西——比如我的薪资、福利，都是重中之重。

What I just said, like my compensation information, my benefits, those are critical.

但想搞清楚这些简直难到爆，我还没见过有谁喜欢这玩意儿。

But navigating that is extremely hard, and I've never met anyone that likes it.

上个月我为了租房子，登录的次数比你多得多，也碰上了一模一样的糟心事：东西到底藏哪儿了？为啥这破系统这么难用？

I had to log in considerably more times than you in the past month because I was applying for an apartment, and I had the exact same pain of, like, where do I find anything and, you know, like like, why is this so difficult to use?

完全感同身受。

So completely completely resonates.

你之前不是还谈到过平台更替吗？这会儿正好有个特别棒的小历史回顾。

I think, like, you also earlier were talking about platform shifts and, like, you have this great kind of mini history lesson right here.

咱们要不要简单聊聊，到底是什么样的平台变迁才催生了Workday？

Like, we should maybe just, like, touch on this for a second, but kind of, like, what exactly was the platform shift that created Workday?

现在咱们也说清楚点，眼下我们正在经历的这个新平台变迁到底是什么？

And, like, now let's, like, just be super clear cut about what this new plat platform shift is that we're seeing right now.

对。

Yeah.

显而易见，我们现在看到的平台转变，就是企业从非AI原生转向AI原生。

I mean, the the obvious answer to the platform shift we're seeing now is is, you know, kind of the shift from businesses that weren't AI native to businesses that are.

这正是机会所在。

That's the opportunity here.

我特别喜欢Workday的故事。

I actually love the Workday story.

它当年可是硬生生把PeopleSoft给吞了。

It was a crazy hostile takeover of PeopleSoft.

有拉里·埃里森，还有这些了不起的创业者，戴夫·杜菲德，然后我得说一下他的姓，布斯里。

There's Larry Ellison, these amazing entrepreneurs, Dave Duffield, then I'm gonna script his last name, Busri.

但你瞧，他们重新出现，说：我们看懂了这个转变。

But you see them come back and say, hey, we see this shift.

我们刚失去了自己的孩子。

We just lost our baby.

我们知道该造什么。

We know exactly what to build.

我们清楚所有那些你得解决的边缘情况。

We know all of the edge cases that you need to solve for.

在2005年解决这些问题，简直难到爆。

And solving for that in 2005 was actually a heck of a lot.

现在做起来也够难的，但那时候更难，毕竟软件开发本身就没现在这么容易。

It's still really hard, but it's even harder back then just considering how hard it was to build software.

所以我觉得这些人真是了不起的创业者。

And so I think these guys are amazing entrepreneurs.

他们看到了这个机会，就把在PeopleSoft刚做出来的东西，重新做成了云原生的版本。

They saw this opportunity and they built the cloud native version of what they had just built at PeopleSoft.

所以他们简直就是完美的创始团队。

And so they were the perfect founding team.

他们打造了这么一家了不起的公司，真的堪称一个了不起的成功故事。

They built this incredible business, and I think it's just an amazing success story.

所以，他们真正写下的，就是这些。

So that's really what they wrote.

当然，中间有个架构上的鸿沟。

Obviously there's the architectural gap.

还有另一个大背景，就是如今几乎所有顶尖的企业软件公司都经历过的一次转变：从‘我们得砸大钱买硬件、堆一堆服务器在后厨’，变成‘我们要的是多租户、能按需付费的系统’。

And then there's this other backdrop that basically every leading enterprise software company of today wrote, which was the shift from, Hey, we're gonna spend a bunch of money on CapEx and have a bunch of servers in some back office to, Hey, we want something that's multi tenant and shifting that into OpEx spend.

他们就是冲着这个方向去设计业务的，也确实做到了。

And so that's what they designed their business to do, and that's what they did.

所以现在回头看，据我所知，还没人做出过功能齐全、能直接用于企业级的人力资源信息系统或HCM系统。

So now when you look at today, there is not, to my knowledge, someone that's built an enterprise ready feature parity HRIS or HCM.

真还没人给企业做过这个。

There's just not someone that's done that for the enterprise.

企业级的。

Enterprise.

我觉得你其实能做到。

I think you can actually do that.

我觉得你能把它做出来。

I think you can build it.

我觉得这难度特别大。

I think it's extremely challenging.

但原因在于，现在企业看到这类软件，会觉得它真能干点事儿——多亏了智能代理让工作方式变了，有了原生AI的软件，人类能做的事就多了。

But I think the reason for that is now companies can look at this piece of software and say, it can do things today because of the way agents allow you to work and the things that humans can do if they have an AI native piece of software.

它现在能做的事，是以前那些版本根本做不到的。

It can do things today that the past versions of this couldn't do.

所以，当你想想大企业要彻底替换核心系统得付出什么代价，那体验必须得完全不同。

And so when you think about what it takes for a large enterprise to rip and replace some piece of core technology, the experience has to be very different.

它的成本结构也得不一样。

The cost profile of it has to be different.

得真能让你省一大笔钱，同时还能促进增长，实实在在地改善员工或客户的体验。

Has to allow you to save a bunch of money, grow, have a better employee or customer experience in a material way.

我觉得这正是我们目前在企业软件其他领域看到的趋势。

And so I think that's what we're seeing with other businesses across enterprise software.

我在前面提到的那些大类里，也都能看到这种现象。

We're seeing that in the other big categories that I mentioned here.

我觉得人力资源领域也会出现这种情况。

And I think we're gonna see it in HR.

所以，这就是机会所在。

And so that's the opportunity.

是的。

Yeah.

为了公平起见，咱们先替那些批评者说句话，因为你刚才提到过，比如说‘我们确实有这个，还有那个’。

And just to maybe straw man the critics for a second because you do talk a little bit about, you know, okay, we have, like, and we have here.

让我找一下，对。

Let me just find Yeah.

对。

Yeah.

找到这儿这一段。

Find this this bit right here.

那你怎么回应那些人呢？他们可能会说：‘我们已经有Deal了，还有Gusto这些公司呢。’

So, like, what would you say what would you say to these guys who are kinda like, well, we have deal and, you know, we have we have Gusto.

我们身边已经有不少初创公司在做这个领域了。

We have all these other startups that are kind of, you know, in the space already.

那到底什么是

Like like, what is

我觉得回答这个问题有两个角度。

So I think there's there's two ways to answer that question.

一个是现在正冲着核心HR和核心薪资系统去的公司，另一个是那些旁支的、相关的机会。

One, there's the people that are going kind of after core HR and core payroll today, and then there's kind of the orthogonal, or adjacent opportunities.

你知道，安德里森显然是Deal最大的投资人，我本人也是Deal的铁粉。

Like, you know, obviously, Andreessen is, I think the biggest investor in deal, and I'm an enormous Deal fan.

我觉得他们团队简直太棒了。

I think it's an incredible team.

但他们做的根本不是这一套。

They're not exactly doing this play.

所以这完全是另一类事情。

So I think that that's just a separate category in and of itself.

我觉得，未来我们在这篇帖子里聊的这种业务，说不定能跟Deal成为绝佳的合作伙伴。

And I think maybe the company that builds this business that we're talking about in this post could be an amazing partner with Deal over time.

所以我觉得这里面有不小的机会。

So I think that there's an opportunity there.

再说到Gusto和其他那些在做全新HR系统的人，也就是我所说的从零开始的策略。

Then you talk about Gusto and the other people that are building what I would call a greenfield type strategy an HR system.

这同样是个机会。

That is an opportunity.

这种现象在很多领域都能看到。

We see that across many different categories.

我们投资的公司里，就有做ERP的，比如Real IT。

We have businesses in our portfolio that are doing that for ERP, like real it.

也有公司正在瞄准CRM这个市场。

There's companies that are going after that for CRM.

其他各个领域也都有企业在这么做。

There are businesses that are doing that in all these other categories.

比如银行业，想想Mercury。

Banking, think of Mercury.

这些都是一些典型的打法。

So there are those plays.

我觉得，只有在平台转型初期，才有可能打造出这样的公司。

I think that that's the only way that you can build these types of companies when it's not in the middle or beginning Or sorry, not the middle, but the beginning of a platform shift.

因为你想啊，如果客户已经用了Workday、ServiceNow或者Salesforce，再想把他们挖过来，简直难如登天。

Because if you think about where you have to go capture these customers, if someone has already become a customer of Workday or already become a customer of ServiceNow or Salesforce, it is so hard to rip them off.

所以这些毛收入留存率才这么高。

That's why these gross dollar retentions are what they are.

它们之所以是黄金标准，是有原因的。

They're the gold standards for a reason.

咱们要不要花点时间看看这张漂亮的图表？

Should we look at this beautiful chart for a second?

对。

Yeah.

这真是个了不起的生意。

It's an amazing business.

我觉得现在的机会是，你可以直接去跟这些客户硬碰硬，不是从零开始，而是从现有客户手里抢，就说：嘿。

And and so what I think the opportunity is now is you can go after these people and go head to head instead of Greenfield, it's brownfield, and say, hey.

你是不是一直被Workday绑着？

Ever you you have been a hostage to Workday.

你一直被困在这儿，只要有人能把新软件做对了，你就能获得自由。

You have been a hostage here, and you can be freed with whatever new piece of software if someone can build this in the right way and form and fashion.

所以这就是机会所在。

So that's the opportunity.

我觉得现在企业内部已经积攒了足够的动力，准备彻底替换旧系统了。

And I think that what we're gonna see is enterprises now have the kinetic kind of energy buildup inside of them to go do the rip and replace.

但你得把东西做对，得聪明地解决这些问题。

But you need to build this in the right way, you need to solve the problems intelligently.

那样的话，你就能做成这件事。

And then I think you'll do this.

肯定有很多人听了这段话，或者看了我的帖子，会说：‘你这纯属吹牛皮吹大了。’

And there's going to be a lot of people who listen to this or read my posts and say, Hey, you're full of hot air.

你根本没认真想过，也没跟谁聊过。

You clearly haven't thought about this or talked to anyone.

实际上，我跟不少财富500强企业，还有不少私营公司的人聊过，他们都在用Workday，我就跟他们说了我心目中该有的产品是什么样。

The reality is I talked to, like, a bunch of Fortune 500 and also private companies that, like, use Workday, and I told them the kind of business that, like, I think should exist.

其中有好几个人直接说：‘要是现在能买到，我立马下单。’

And some of them were like, I would buy this from you if it existed right now.

所以这件事其实已经在发生了。

So it is actually already happening.

没人喜欢在Workday里做这些乱七八糟的任务。

No one likes doing these crazy tasks inside of Workday.

员工都不愿意跟Workday打交道。

No employee likes interacting with Workday.

谁都不想登录这个门户。

No one wants to go into this portal.

所以如果你能出去解决这些问题，那就有巨大的市场等着你去抢。

And so if you can go out there and solve these problems, there's a huge pie to go after.

在我们深入讨论这些问题到底是什么、机会在哪之前，我觉得咱们得先说说Workday现在在干嘛——他们可没闲着。

So before before we get, like, super granular about what these problems are and what the opportunity space is, like, I feel like we should talk about what Workday is doing right now because, like like, it's not like they've been sitting on their asses.

是的。

Yeah.

他们的AI年经常性收入已经达到了四亿美元。

Their AI ARR has hit, like, $400,000,000 in yearly run rate.

每年增长都是三位数。

It's growing triple digits year on year.

他们可没闲着。

Like like, they're they're not doing nothing.

你能聊聊他们现在有哪些产品吗？你觉得哪些地方做得不够，或者哪些地方其实还不错？

So can you talk about the products that they do have and kind of where you feel like they fall short or maybe where you feel like they're sufficient or good?

对。

Yeah.

我写这篇文章本来就想保持平衡，因为我觉得——我之前也说过，

So first, like, I I wrote this piece to be balanced because I do think, like again, I said it.

Workday 是我见过最棒的软件公司故事之一。

Workday is one of my favorite software stories ever.

这个生意的护城河简直太深了。

Like, I think the business is unbelievably defensible.

直到一年前，我压根儿没想过要写这么一篇文章。

Until twelve months ago, I would have never thought about even writing this post.

我花了整整十二个月才下定决心写这篇文章，因为得先说服自己这事真有可能。

And it took twelve months of me thinking about this to write this piece because I had to convince myself that it was possible.

我觉得我看到的，是那些裂缝的早期迹象。

I think what I'm seeing are the early signs of cracks.

我特别想说的一点是——这不光是Workday的问题，现在太多企业软件公司都在吹自己的AI收入指标，简直太离谱了。

And so one of my favorite parts of this, and this is not just specific to Workday, but there are so many different enterprise software companies now that are reporting some metric affiliated with AI revenue that I think is just kind of ridiculous.

我跟专业人士聊起flex credits的时候，一提这词，人家就笑场。

And so when I talked to professionals about the flex credits, I would say flex credits and people would just start laughing.

这种反应挺有意思的，其实我写这篇文章，就是想说这个。

And it's interesting to get that reaction because I it's basically I wrote this.

说白了，这不过是个采购方式上的小花招。

I think it's basically just a procurement innovation.

我觉得根本没什么实质性的变化。

I don't think anything is actually fundamentally changing.

买个Workday Extend的授权，再开发个应用，还是得花两万五千美金。

It still costs $25,000 to go get a Workday Extend license and then build an app.

这些灵活积分本质上还是采购层面的创新。

These flex credits are still just a procurement innovation.

在人们的Workday系统里，并没有真正出现什么智能代理型体验。

There's not actually agentic experiences that are being delivered inside people's Workday instances and Workday platforms.

现在最大的笑话就是，所谓的AI收入根本不是真正的AI收入，不过是几个CIO互相说：‘我得报点AI收入’而已。

So I think that that's a big kind of farce right now is that this AI revenue is actually AI revenue instead of just one CIO talking to another saying, I need AI revenue.

我得在AI上花钱。

I need to spend money on AI.

咱们凑一块儿吧。

Let's get together.

你懂的。

You know?

我开玩笑说，咱去Ruth's Chris，把所有东西都卖了。

And I joke, let's go to Ruth's Chris and sell everything.

你知道的，没什么比一块黄油蛋糕加采购创新更能搞定这事了。

Like, you know, nothing like a butter cake and, procurement, innovation to, to make it work.

是的

Yeah.

对

Yeah.

我现在觉得，这让我们不得不面对一个问题：

I and and now I think this brings us to kind of like, okay.

你知道吗？这正是这篇文章后半部分的标题所要探讨的：

What you know, I this is what the title of, you know, the entire second half of the piece is.

一个原生AI的Workday应该是什么样子。

What an AI native workday should look like.

没错。

Yeah.

咱们不如直接聊聊这几个方面吧，你见过哪些案例？听到过哪些说法？接触过哪些创始人？还有那些财富500强企业到底在提什么需求？说白了，这东西到底该是什么样？

Should we should we just kind of talk through a couple of these categories and kind of what what you've seen, what you've heard, the kind of founders that you're talking to, also the kind of demands that you're hearing from Fortune five hundreds and just kind of, like, what does this look like?

对。

Yeah.

我觉得写这个是因为这六个特性真的特别关键。

I I think, like, I write this because I think that these are the six very important properties.

然后我会简单聊聊，具体该从哪儿开始下手。

And then I I talk a little bit about, like, tactically how I think you should start.

对。

Mhmm.

不过我觉得，大家迟迟不换的根本原因，主要就在这几点。

I think that these were the big ones, though, of, like, why do people not switch?

对吧？

Right?

那到底要怎么才能让人愿意换呢？

And, like, what could you get what could get someone to switch?

我觉得第一个点，现在只要是见过AI代理能做什么、见过那些纯AI原生的迁移或实施公司怎么干的，大家都会认同。

I think the first one is, like, unanimously understood now to anyone who's, like, seen what coding agents can do and has, like, seen what, like, migration businesses or, like, implementation businesses that AI native are doing today.

我觉得这种系统，三十到六十天就能部署起来。

I think you can deploy a system like this in, call it, thirty to sixty days.

而过去，这通常得花上一年以上，甚至更久，还得多花一大笔钱。

Whereas historically, this is twelve plus months easily, and can be much longer than that and cost you a bunch of money.

这种情况一定会发生。

That is going to happen.

整个企业软件领域都会这样。

It's going to happen across enterprise software.

所以这才是为什么现在这些业务能被轻易取代和重构的根本原因。

And that's why that is the fundamental building block of why these businesses now can be ripped and replaced much simpler.

我提到 Workbench Native 的时候，其实想了很久该叫它什么。

When I talk about Workbench Native, I thought a lot about what to call this.

但如果你去问 HR 管理员，他们每天得干一堆乱七八糟的事儿才能让 Workday 正常运转。

But if you talk to an HR administrator, there's so many random things that they have to do to make Workday work on a daily basis.

比如公司里某些岗位的薪资计算参数不一样，他们就得用 Google 表格手动处理。

This can be something as simple as doing payroll for a certain type of role inside the company that doesn't have exact same parameters, and so they have to do it on a G sheet.

也可能是要新建一个业务单元，得做一堆配置。

It could be spinning up a new business unit and doing a bunch of configuration.

比如考虑员工体验，Elena，你最近才从纽约搬过来。

It can be trying to think about an employee experience where Elena, you lived in New York until recently.

你刚到旧金山的时候，就想问问：嘿，谁在旧金山啊？

You were coming to San Francisco and you wanted to know, Hey, who's in San Francisco?

Joe在旧金山吗？

Is Joe in San Francisco?

我该怎么查到这些信息？

How do I actually find out that information?

在Workday里，根本找不到靠谱的方法。

There's no good way of doing that inside Workday.

那该怎么打造这样的体验呢？

So how do you build that experience?

所以我觉得，Workbench Native 的意义就在于，让HR团队能自己动手搭建，而不是花两万五千美金请人做定制，再找个顾问来帮你搞。

So I think that this Workbench native, allowing HR teams to genuinely build versus, hey, spend $25,000 to build anything on top of Workday, and then also get yourself a consultant that can build it for you.

我觉得，企业软件里那种老一套的做法，已经过时了。

I think the days of that in enterprise software are gone.

任何试过用云代码开发的人，都会告诉你，这种体验迟早会普及开来。

Anyone who's experienced the joy and magic of building things with Cloud Code is going to tell you that is going to be everywhere at some point.

我觉得这会非常关键。

So I think that's gonna be really critical.

这种以智能代理为核心的思路，想想我之前说的，或者你遇到的那些想从Workday里捞信息的麻烦事——未来所有系统都得这么干。

This agent first piece is I think that thinking about the way that I talked through my issue or your issue of trying to get anything out of Workday, I just think that agent first is going to be the way that these systems work in the future.

你肯定看到了Salesforce的无头架构公告。

You saw the Salesforce headless announcement.

他们正拼命往这个方向走，Workday其实也在这么做。

They're doing as much as they possibly can to shift towards this future, and I see Workday doing this too.

未来大家会直接在工作工具里搞定一切，或者出现一个原生AI版本，让智能代理替你干活。

So think that people are gonna stay inside the work the tools that they're using, or there's gonna be an AI native version of this so you can just actually allow agents to do work for you.

我觉得开放性特别重要。

I think the open piece is critical.

安全是底线，权限管理也绝对不能马虎。

Security is paramount, and permissioning is really critical.

然后是合规问题，这是最后一点，因为这个领域监管非常严格。

And then the compliance piece is the last one, which is this is a heavily regulated space.

所以从一开始就得想清楚，怎么监控、怎么确保你的平台能满足企业级和全球化的需求，这非常关键。

And so just thinking through how do I monitor that and how do I ensure that my platform is enterprise ready, global ready from the start is critical.

总之，我觉得这些就是零阶段产品该有的样子。

So anyway, I think that's the properties exactly what it looks like in day zero though.

我觉得一些优秀的创业者得想办法，用他们最初的那个工具，给买家制造那种动力和紧迫感。

I think that some great entrepreneurs are gonna have to figure out how to generate that kinetic energy and that urgency for the buyer with whatever that initial tool is.

对。

Yep.

咱们再聊聊最后这部分吧。

And then let's just talk about the end of the piece a little bit.

你其实是在说，这事儿不光是HR软件的事儿。

So you sort of argue that, you know, this is this is actually about more than, you know, HR software.

更深层的是，人和智能代理该怎么互动。

It's also about, you know, how humans and agents interact.

我真想再听听你详细说说这一点，还有这里。

And I I would love to just hear hear you talk a little bit more about that and here.

这里。

Here.

对。

Yeah.

在这儿。

Down.

是的。

Yeah.

我们就在这儿。

We're down here.

我真的想再多听听你讲讲这个，你到底指的是什么？

Like like, I I would love to hear a little bit more about that and kind of, you know, what do you mean when you say that?

你能大概描绘一下这会是什么样子吗？

And and kind of can you help sketch out what that looks like?

是的。

Yeah.

我的意思是，随着越来越多的智能代理替人类做事，搞清楚它们在公司里的角色和权限会变得特别关键。

I I mean, I think that, like, as you have more agents doing work on behalf of humans, like, understanding their roles and permission inside of the company is going to be critical.

那你怎么利用企业里最重要的一批数据，来指导这些智能代理的未来发展呢？

And so how do you actually use what is one of the most critical pieces of or corpuses of data inside the enterprise to inform that agentic future?

在我看来，你得先有个核心代理来做这件事。

In my opinion, you need an agent first kind of core to do that.

我觉得这会是CIO们的一大难题。

And so I think that's going to be a big problem for CIOs.

我确实听不少大公司的CIO说过，比如：

And I've heard this actually from many big CIOs of just like, alright.

我们该怎么管权限？

Like, how do we do permissioning?

随着我们用越来越多的工具，又该怎么追踪和管理呢？

How do we do tracking across our company as we adopt more tools?

所以我觉得这可能就是整个背景。

And so I think that could actually be like a backdrop of this.

这意味着，我们用的那些第三方工具，不只是自己开发的，也都需要访问这些数据，而且得是干净、统一、原生支持AI的方式。

That could also mean, hey, all these other third party tools that we're adopting, not just building, also need access to this information, and they need it in a clean and cohesive way that's kind of AI native.

所以我觉得这会是另一个推动力。

And so I think that's going to be another driver.

我们其实已经在IT服务管理领域看到了，这跟人力资源关系很紧密。

I think we're seeing this already in ITSM, which is tightly related to HR.

所以说，我们正看到世界上那些最稳固的企业的裂缝正在出现。

And so I think, again, we're seeing the cracks in the most defensible you know, businesses in the world.

这真是个让人兴奋的时刻，比赛已经开始了。

And and it's a really exciting time, and and the race is on.

这就是为什么Workday把Neil请回来的原因。

Like, this is why Workday brought a Neil back.

这就是为什么他们已经做了两轮裁员。

Like, this is why they've, you know, done two layoffs.

这就是为什么他们拼死拼活，像收购Sana这样的公司，就是为了抢在前面——这是一场竞赛，赢了的回报太大了。

This is why they're, like, fighting tooth and claw and and, you know, buying businesses like Sana, you know, to try to to to find this off because it's a race, the prize is huge.

那你会不会说，HR软件其实是软件领域里最能反映人类社会现状的垂直领域？

So so would you would you make the argument that HR software is, like, the most anthropologically salient sort of vertical in software in terms of telling us where we are.

它能告诉我们，我们正处于软件发展的哪个阶段。

Like, it tells us where we are in the software cycle.

它也能反映出我们当下的社会状态。

It tells us where we are, you know, socially.

普通人普遍都期待有这些福利。

Like, you know, people on average expect these kinds of benefits.

他们习惯了每年的健康保险、会员服务，还有各种其他福利。

They like their yearly function health membership and their, you know, whatever else.

这些在职场里早就习以为常了。

And, you know, like, this is acceptable in the workplace.

但那种就不行。

This is not.

现在到了AI时代，不仅会带来全新的平台变革，还会彻底改变采购方式和社会规范，方方面面都会不一样。

And now, like, the AI era is going to just introduce, you know, both a completely new platform shift but also completely new buying and procurement norms, completely new social norms, all of these things.

如果你想看社会接下来会往哪儿走，就看看HR软件吧。

And if you wanna see where society is going, you should look at HR software.

我

I

我也这么觉得。

think so.

说实话，我基本同意你所说的全部观点。

I mean, you know, I agree with basically all of that.

要说历史上有点疯狂的是，HR部门从来都不是公司里最敢试新东西、最积极跟上技术节奏的那个部门。

I think the crazy maybe the the crazy part about it historically is that this has not been like the the front foot of the organization, you know, like, least in terms of, like, you know, trying new things or adoption cycles.

所以我觉得，也许AI真正开始大规模普及的时候，第一个信号就会出现在HR领域。

And so I think that we'll, like maybe it will be the it'll be like the beacon of AI is actually finally reaching more mass kind of take off.

因为对我来说，这恰恰反映了我们当下所处的阶段。

Because that to me feels like part of where we are right now.

这就像是你手里有一张美国地图，沿海地区或者大城市里亮着灯——你懂我的意思吧。

It's like if I had a map of The United States and there's lights in the coasts or I don't know how you'd say it, like in the major cities.

那些亮着的灯，代表那些已经发现AI能彻底改变业务流程的人，他们特别兴奋。

And it's like the lights indicate who has discovered that AI can actually dramatically transform business processes, and they're excited by it.

但我觉得，全国其他地方还都是黑漆漆的。

And I think the rest of the country is kind of dark.

我觉得未来十二到二十四个月，这张地图会突然亮起来，变化特别明显。

And I think what we're gonna see over the next twelve, twenty four months is just a dramatic brightening of that map.

等这种情况发生，人力资源部门就会跟上，那时候才是真正起飞的时刻。

I think as that happens, like, we're gonna see HR start to swing, and that will be like when we really start to see takeoff.

这么说，可能和你刚才说的其实是同一个意思。

So that maybe is a different way of saying the same thing you just said.

对。

Yeah.

没错。

Yeah.

是的，人人都会用上原生AI的HR工具。

Yeah, everybody gets their AI native HR tools.

所以我觉得，咱们最后一个问题吧，你刚写完的那篇是关于Revit的。

So I think I think for for our final question, so the last piece you wrote was on Revit.

对。

Yes.

这儿的每个p都……

The every every p here.

咱们看看这个会不会……嗯，行吧。

Let's see if this is gonna yeah.

是的。

Yeah.

你去过的每一栋楼，都是用1997年开发的软件设计的——回头想想，这话确实有点儿炸裂。

Every building you've ever been in was designed by software built in 1997, which in retrospect was a bit of a controversial statement.

当时不少人不买账。

People didn't like that.

他们说：等等。

They were like, wait.

97年之前可早就盖过楼了。

There have been buildings built before '97.

这个我当然知道。

I'm like, I'm aware.

我找房子的时候，有个硬性要求：必须是1997年以后建的，因为我压根不信以前的东西。

That was actually that was a prerequisite for me when I was looking for apartments that had to have been built in 1997 or later just because I I didn't I didn't trust anything.

97年之前的我全都不信。

I didn't trust anything prior to that.

但我们提到了一首歌，可能是有史以来最伟大的专辑里最伟大的一首歌。

But we we had a reference to maybe one of the greatest album one of the greatest songs on the greatest albums of all time.

《Once in a Lifetime》，来自Talking Heads的专辑《Remain in Light》。

So once in a lifetime off of the album remain in light by the talking heads.

经典专辑。

Great album.