Arca 如何让长跨度智能体驱动金融服务业务

简要说明Brief Description

Aman Singh 介绍 Arca 如何利用信号、运营上下文、持久记忆和共享空间，让智能体能够跨越数日推进金融服务相关工作。Aman Singh explains how Arca uses signals, operational context, persistent memory and shared spaces to let agents drive financial-services work across days.

目录Table of Contents

• 简介：Arca，一家原生 AI 服务公司
  Intro: Arca, an AI-native services firm
• 长时程工作与长时间运行的工作
  Long-horizon vs long-running work
• 示例：为客户筹集现金以支付税款的情形
  Example: raising cash for a client's tax bill
• 代理端到端地负责结果
  Agents own the outcome end to end
• 追踪一个信号：一封电子邮件的旅程
  Following one signal: the journey of an email
• 空间：智能体吸引人们加入的地方
  Spaces: where agents pull in people
• 智能体观察世界，让人们不必亲自操心。
  Agents watch the world so people don't have to
• 应对纷乱复杂的现实世界
  Handling a messy real world
• 长周期工作的基本要素
  The primitives of long-horizon work
• 涌现的复杂性，以及过去六个月
  Emergent complexity, and the last six months
• 为什么要这样做？智能体负责运营业务。
  Why do this? Agents run the business
• 通过引导智能体来引导业务
  Steering the business by steering agents
• 使用 AI 工具与原生融入 AI工作流的区别
  Using AI tools vs being AI native

简介：Arca，一家原生 AI 服务公司Intro: Arca, an AI-native services firm

我叫阿曼，在 Arca 负责 AI。今天我想和大家聊聊 Arca 的长期运行智能体，它们负责运营我们的业务。Arca 是一家 AI 原生服务公司。我们为客户提供金融服务，例如财务规划、税务规划、资产管理和投资建议。我们的客户群体多种多样。My name's Aman. I lead AI at Arca. and today I'm going to talk to you about Arca's long horizon agents that run our business. Arca is an AI native services firm. We provide our clients with financial services like financial planning, tax planning, asset management, investment advice. and we serve a range of clients.

他们中的一些人就像科技行业的高管，积累了大量财富，需要一份切实可行的提前退休计划。但也包括新罕布什尔州的那个家庭，他们刚赚到人生第一个百万美元，终于想买下那栋房子。我说过，我们是 AI 原生的。如今这个术语有很多种定义，而且已经被过度使用了。Some of them are like technology executives that have amassed a lot of wealth and need a real plan to retire early. But also that family in New Hampshire that made their first million dollars and want to buy that house finally. I said we were AI native. There are lots of definitions of this term these days; it is very overused.

但我们所说的 AI 原生，是指由智能体来运营我们的业务，而不是让它们躲在某个副驾驶的角落里协助业务。But what we mean by AI native is that agents run our business. Not that they assist the business from some co-pilot corner.

他们执行那些决定我们业务为客户带来何种成果的决策。They execute the decisions that shape the outcomes that our business has for our clients.

长时间跨度的工作与长时间运行的工作Long-horizon vs long-running work

在这个行业工作，了解它是什么样子很有意思。在金融服务领域，工作周期更长，但连续运行的时间更短。我的意思是，这里不像是智能体连续数小时甚至数天不停地消耗 token，对吧？更多是围绕一个客户目标展开，而这个目标需要 10 天才能完成。整个过程时断时续，中间有很多参与者、许多决策点和多次交接，最终这个目标大约在 10 天后才达成。这就是我们正在处理的这类工作。Work in this industry, it's interesting to know what it looks like. In financial services, the work is more long horizon and less long running. And what I mean by that is this is less agents grinding tokens for like hours or days on end, right? This is more one client goal that spans 10 days. It's very stop-start. There's many actors in between, many decision points, many hand-offs. and ultimately that goal lands like 10 days later. That is the kind of work that we're working on.

示例：为客户筹集现金以缴纳税款Example: raising cash for a client's tax bill

下面是一个例子。简是我们的客户，经济状况相当宽裕，但在 4 月 15 日要缴一大笔税款，对吧？所以，我们会帮简找出筹集这笔钱来缴税的最优税务方案。顾问会负责分析所有账单，并研究如何实现税务优化。And so here's an example. Jane's our client, fairly wealthy, but has a big tax bill on April 15th, right? So, we're going to help Jane figure out what is the most tax optimal way for this cash to be to be raised to pay the tax bill. So, there's the advisor that will do all of the statement analysis and the analysis of how to make this tax optimal.

有交易员会实际负责清算这些资产。资产本身需要好几天才能结算，之后运营人员会介入，触发 ACH 转账。转账到账本身也要好几天。所以，这就是我在这个行业里说的这类工作——需要好几天才能结算。There's the trader that would actually liquidate the assets. The assets themselves will take like days to settle, and then the ops people will come in and kind of like trigger the ACH. That will itself take days to land. So, this is kind of the work I'm I'm talking about in this industry, which takes like days to settle.

当然，如果你为多个客户大规模开展这类工作，我们事务所完成这类工作的时间就会呈指数级增长，对吧？And of course, if you do this at scale for several clients, the amount of time our firm takes to do this sort of work exponentiates, right?

代理负责端到端地交付成果Agents own the outcome end to end

所以，在 Arca，我们采取了不同的方法。了解 Arca 并不是一家 SaaS 公司，这一点很有帮助。我们不是把软件卖给其他企业或公司，我们自己就是这家公司。我们是一家应用型 AI 公司，只不过恰好在这个行业经营服务业务。因此，我们采取的方法是：好，我们要让智能体端到端地负责整个长周期的成果。So, at Arca, we took a different approach. It's helpful to know that Arca is not a SaaS company. We are not selling software to other businesses, to other firms. We are the firm. We're an applied AI company that happens to be running a services business in this industry. And so, the approach we took was, all right, we're going to have agents own this whole long horizon outcome end to end.

为此，我们给了他们几个关键组件，先让你一睹为快。我们为他们提供来自外部世界的信号，让他们能充分了解正在发生的事情。因此，智能体会读取这封邮件。我们还向他们提供了所有运营背景信息，告诉他们该如何处理这封邮件。这个行业通常如何应对这种税务情况？And for that, we gave them a couple of key components, I'll give you a sneak peek of. We gave them signals from the external world that keeps them really aware of what is happening. So, that email will be read by agents. We give them all of our operational context on what to do about that email. What does this industry do with this situation of taxes?

对吧？我们给他们提供了一个运行框架，配备了所有工具和防护机制；更重要的是，它能在较长的时间跨度内持续保留记忆，帮助他们完成这项工作。Right? We gave them a harness that gives them all of the tools and the guardrails, and more importantly, persistence of memory over the kind of long horizon period to carry that work out.

当然，我们也赋予了他们与人协作的能力。因此，与智能体协同完成的工作周期大大缩短了，因为我们所做的，只是让人类不再处于这项工作的关键路径上。没人需要连续几天盯着这个目标不放，智能体会负责推进。在这次演讲中，我很乐意向大家介绍我们究竟是如何做到这一点的。And of course, we gave them the ability to collaborate with people. So, that work with agents compresses a lot because all we did was we took humans out of the critical path of this work. Nobody has to babysit this goal over days. Agents are on it. In this talk, I'd love to describe to you how we actually do that.

追踪一个信号：一封电子邮件的旅程Following one signal: the journey of an email

我们来跟踪这封邮件的旅程。简需要为这10万美元缴税。这封邮件和我们平台上的许多其他信号一样，会被、会被后台代理读取。所以，这里有我们公司打出的电话，有收到的文件，有举行的会议，有托管机构发来的事件，还有正在发生的市场变动。Let's follow the journey of this email. Jane owes on the $100,000 for taxes. This email, much like many other signals in our in our platform, will be will be read by background agents. So, there's like phone calls our firm is making, there's documents that are coming in, there's meetings being had, there's events coming in from our custodian, there's market movement that's happening.

这些都是我们的后台智能体为了推进工作而持续评估的信号。所以，这封邮件就是其中一种信号。而当信号是由智能体发出时，它们就会真正开始处理工作，对吧？我们有一套执行框架，能让它们了解该如何应对当前情况。它们对客户有深入的记忆，所以知道：“嘿，千万别动 Jane 的苹果持仓，因为她大概 8 个月前在一次通话里说过这件事，对吧？”These are all signals that our background agents are continuously evaluating to progress work. So, this email is one of those signals. And when the signal is by agents, they actually get to the work, right? we have a harness that gives them the knowledge to know what to do about the situation. They have deep client memory, so they know, "Hey, let's never touch Jane's Apple position cuz she said that in a call like 8 months ago, right?"

当然，我们有一个沙箱，用于执行这项分析所需的所有财务计算。但总的来说，我认为如今的各类执行框架正逐渐趋同于一种标准形态。因此，更值得关注的是执行框架之外会发生什么。完成工作后，它们会生成这份计划。we of course have a sandbox for all of the financial computation that needs to happen for this analysis. but largely, I think harnesses today are coalescing on a standard shape. So, it's more interesting what happens outside of outside of this harness. So, after they've done the work, they produce this plan.

Spaces：智能体吸引人们参与的地方Spaces: where agents pull in people

这个计划本身无法在没有人的情况下执行。所以在我们的产品中，我们会创建一个空间，智能体也会创建一个空间，把计划发布出来。你可以和智能体聊天，了解计划的进展情况，而且可以有多个人同时和智能体聊天。This plan by itself cannot be executed without people. So, in our product, we build a space, agents build a space, where they put out this plan, and there's you can chat with the agent to like kind of know what's happening with the plan. There's multiple people that can be chatting with the agent.

更重要的是，当他们需要人手来执行某些工作时，也会在这个空间里把人拉进来。所以在这个案例中，我们会批准计划，然后给 Jane 发邮件。接下来就会有一整套来回沟通：Jane 回复了吗？她的……她的回复对我们原先的计划意味着什么，对吧？And more importantly, this space is where also they pull people in when they need people to execute some work. So here in this case, we would approve the plan and the email out to Jane. And then there's this whole dance of has Jane replied? What does her What does her answer mean for the plan that we had, right?

然后，在整个漫长的工作过程中，还必须回答所有这些其他问题。And then all of the these other questions that will have to be answered throughout this whole long-horizon work.

智能体观察世界，让人们不必亲自操心。Agents watch the world so people don't have to

在 Arca，我们当时想，当然不能让人一直盯着它。所以我们让智能体替我们观察这个世界，这样人就不用操心了，对吧？我们为它们提供工具，让它们能够登记自己对我们世界里发生的任何事情的关注。比如，它们可以说：“Jane 回复的时候通知我。”“这笔 ACH 到账的时候通知我。”对吧？这背后由我们的信号系统提供支持，智能体会把信号路由给对此感兴趣的其他智能体。当然，如果信号一直没到，智能体也可以按计划自行唤醒，以便持续跟进工作。At Arca, we were like, of course we're not going to have people babysitting it. So we let our agents watch the world. So people don't have to, right? So we give them tools to be able to register interest in anything happening in our world. So they're able to say, let me know when Jane responds. Let me know when that ACH lands. Right? And this is powered by our signal system in the in the background where agents are routing signals to agents that are interested in it. And of course, if the signal never lands, agents can like wake themselves up on a schedule as well to keep track of work.

应对混乱的现实世界Handling a messy real world

但现实世界当然非常混乱，对吧？有人在等那封邮件，但 Jane 始终没有回复。假设我们当面见到了她，并了解到一些会影响计划的信息。于是，有人进入我们的界面，把这些信息添加进去。但添加信息或在界面中进行操作这一行为，本身也是一个信号，会唤醒某个智能体。But then, of course, the real world is very messy, right? People are waiting for that email, but Jane never responds to that email. Say we meet her in person and we learn some information that affects the plan. So the person goes in and they add this information into our UI. But this act of adding information or doing work in our UI is also a signal that will wake up an agent.

事实上，当人们在我们的产品中工作时，比如重新安排日历事件，这就是一种信号；将任务标记为已完成，也是一个信号。因此，这套系统让我们能够同时了解外部世界和内部世界中随着工作推进而发生的一切，从而让我们的智能体能够参与到工作中。In fact, when people work in our in our product, when they reschedule a calendar event, you know, that's a signal. When they mark a task complete, that is a signal. So this system lets us be both aware of everything happening in the external world and the internal world as work is progressing so our agents can participate in work.

当然，智能体之间也会相互通信。因此，后台智能体读取到这个信号后，就能通知另一个智能体更新空间里的信息：“嘿，我们现在只需要 8 万美元，而不是 10 万美元。”And of course, agents are communicating with each other as well. So, this signal read by a background agent will let it notify another agent about this so that it can update the space to say, "Hey, we just need $80,000 instead of the $100,000."

因此，由于这些信号，工作内容发生了变化。故事的最后一部分是，智能体能够在恰当的时机引入合适的人。左侧的智能体读取了我们的运营背景，并判断：“好，这里的下一步是请负责交易这些资产的人参与进来。”于是，这个人现在接到了一项任务，并被邀请参与这项长期工作。So, the work has changed because of because of these signals. And the last piece of the story is that agents are able to pull in the right people at the right time. So, the agent on the left has read our operational context to be like, "Okay, this the next step here is to pull in the person that trades these assets." And so, that person now gets a task and gets pulled in to participate in this long-horizon work.

长周期工作的基本要素The primitives of long-horizon work

总的来说，如果退一步来看，我们为智能体提供了推动长期工作的基础能力。人们在工作中会做什么，对吧？他们需要持续了解正在发生什么，所以我们为智能体构建了一套非常强大的信号系统，让它们能够及时掌握情况。他们需要知道接下来该做什么，所以我们为它们提供了所有这些运营上下文。他们需要实际完成工作，所以我们提供了执行框架，让它们能够真正开展工作并跟踪进展，对吧？还有能够长期持久保存状态的记忆系统，以及在合适的时机引入合适的人。What we've done overall, if you take a step back, is we've given our agents the primitives for driving long-horizon work. What do people do at work, right? They need to stay aware of what is happening, so we've built this very intense like signal system for our agents to be aware. They need to know what to do next, so we've given them all of this operational context. They need to do the work, so it's our harness that lets them actually perform the work and keep track of it, right? It's long-persistence memory systems that hold the state. And bringing in the right people at the right time.

涌现复杂性与过去六个月Emergent complexity, and the last six months

乍一看，这个想法相当简单，但其中蕴含着许多涌现出的复杂性。你可以想象，在任何时刻都会有数百个这样的信号涌入；任务不断被加到人们的待办事项中，又被人接手，后台代理之间也在相互通信，对吧？所以，过去六个月里，我们一直在努力把这个系统打磨得更加出色。This is a pretty simple idea on the face of it, but there's a lot of emergent complexity. You can imagine at any point there is hundreds of these signals coming in, and they are, you know, tasks are being added on people's plates, they're being pulled in, background agents are communicating with each other, right? So, over the last 6 months, we've we've been working on getting this system to be really good.

我们一直在努力填补所有上下文缺口，让智能体充分了解情况。我们也在提高任务分配的精准度，确保在合适的时间找到合适的人。需要留待将来跟进的工作会记录在智能体的记忆中。现在就能完成的工作则会变成一项任务，对吧？此外，我们也一直在提升长期保持状态的能力。We've been closing all of our context gaps so that agents are fully aware. We have been working on the precision of tasks, so pulling in the right person at the right time. Work that needs to be tracked for the future is tracked in the agent's memory. Work that is doable now becomes a task, right? So, we've been working on our ability to hold state over long horizons as well.

而且，因为我们本身就是一家事务所，而不只是一家 SaaS 公司，我们一直在改变并标准化事务所内部的工作方式，并将这些背景信息落实到智能体的状态中。And because we are the firm and not just a SaaS company, we've been changing and standardizing how work happens in the firm in the first place and operationalizing that context into our agent's state.

而且，由于使用我们产品的是我们的同事，我们会收集他们丰富的推理过程，并将其转化为评估项，以改进我们的系统。And because these are our colleagues that use our product, we collect rich reasonings from them that become evals for our system to improve.

为什么要这么做？智能体负责运营业务。Why do this? Agents run the business

所以，我想这里有个问题值得问问：为什么要这么做，对吧？这能解锁什么？这里的核心启示是什么，对吧？对我们来说，这就是我们的整个理念：我们希望我们的平台能够如此原生地融入 AI，以至于由智能体来推动业务发展。智能体负责运营业务，也由它们做出影响我们业务成果的决策。So, I guess there's there's a question to be asked here, like why do this, right? What does this unlock? What is the takeaway here, right? For us, this is our whole thesis. Is that we want our platform to be so AI native that agents drive the business. Agents run the business. They are making the decisions that shape our outcomes.

因此，在 Arca，人们登录后会看到一份由智能体为他们准备好的任务清单。他们已经读过那封邮件，也看过刚收到的文档，差不多完成了 90% 的工作，只在需要你采取行动时才让你介入。构建这套系统，让我们能够在 Arca 实现这样一种状态：智能体发起请求、推动人们行动的次数多于人们向智能体发出请求，从而让我们的智能体推动公司开展的工作。And so, at Arca, when people log in, right, they see a list of tasks that have been prepared for them by agents. They've already read that email. They've like looked at the document that came in. They've done like 90% of the work and they're pulling you in just when it's time for you to act. So, building this system lets us get to a point where at Arca, agents prompt people more than people are prompting agents, which lets our agents drive the work that our firm does.

由于智能体正在推动公司的工作，因此也能推动公司的业务发展。And because agents are driving the work the firm does, agents can drive the business.

通过引导智能体来引导业务Steering the business by steering agents

我举个例子。还记得简发给我们的那封邮件吗？这算是有点失败，对吧？因为一开始我们本来就应该主动处理这件事，对吧？所以，简根本就不该需要发这封邮件。比如说，在我们的 AI 原生业务中，我们确立了一个新目标。我们的新目标就是主动性。Let me give you an example. Remember that email that Jane sent us? This is a bit of a failure, right? Because we should have just been proactive about this in the first place, right? So, Jane should never have to send this email. So, say at you know, our AI native business, we take on this new goal. Our new goal is proactivity.

每年一月，主动整理好每位客户的税务情况，并在他们开口询问之前就联系他们。这是我们业务要推动的一项新目标。而在一家原生 AI 公司里，我们只需要把这项目标纳入智能体的目标体系中。由于它们具备推动员工为业务开展工作的基本能力，便会开始推动这项目标的实现。就这样，到了每年一月，这些分析就会完成，而每位客户的完整方案也会通过一键审批邮件呈现在用户面前。Every January, prepare each client's tax situation and proactively reach out to them before anyone even asks. This is a new business goal for our business to drive. And in an AI native firm, all we have to do is make this goal part of our agents' set of goals. And because they have the primitives to be able to drive work through people for the business, they start driving this goal, and just like that, you have you have these analyses being done in January, and these one-click approval emails with the full plan for every client showing up on our users' plates.

而且，由于它们与主动性的目标一致，因此可以一键审批，推动业务实现这一新成果。And because they're aligned with the goal of proactivity, they can one-click approve and drive the business towards this new outcome.

使用 AI 工具与原生采用 AI 的区别Using AI tools vs being AI native

所以，我认为这大概就是使用 AI 工具的公司与 AI 原生公司之间的关键区别。在一家 AI 原生公司里，当你能够通过引导智能体来引导业务时，你就知道自己已经打造出了一家这样的公司。这也是我们在 Arca 的目标，而且我们正在招聘。所以，请访问我们的网站。如果你有兴趣参与解决这类问题，欢迎联系我。感谢聆听。So, this is sort of I think the key difference between a company that uses AI tools versus a company that is AI native. In an AI native company, you know you've built one when you can steer your business by steering your agents. And that is our ambition at Arca, and we are we're hiring. so, please check out our website. Please reach out to me if this is something that you're you're interested in working on these sorts of problems. And thank you for listening.