一辆运输车驶入利雅得郊外的一个建筑工地，卸下一批散装水泥，并通过 WhatsApp 发送了一张送货单的照片。在 BRKZ 发展的大部分时间里，这张照片是后续一系列人工工作的起点。过去，必须有人手动处理这些信息。如今，我们称之为 Nusa 的智能代理承担了大部分工作。它能读取图像、提取详细信息、将其与我们的系统进行匹配、核实数量并完成交付确认。
A transporter pulls up to a construction site outside Riyadh, unloads a delivery of bulk cement, and sends a photo of the delivery note through WhatsApp. For most of BRKZ’s history, that photo was the start of a chain of human work. Someone had to Today, an agent we call Nusa does most of that. It reads the image, extracts the details, matches them against our systems, verifies the quantities and closes the delivery. The team handles uncertain matches and exceptions.
团队仅负责处理不确定的匹配项和异常情况。目前，Nusa 已经完成了我们对账流程中大部分的交付确认工作，且其处理份额每月都在增长。
Nusa now closes the majority of deliveries in our reconciliation workflow, and its share grows every month. Bulk cement isn’t a glamorous AI use case.
散装水泥并不是人工智能应用中光鲜亮丽的案例。但这恰恰是我喜欢它的原因。
That’s exactly why I like it. I recently came back from a trip to the US where AI dominated most of my conversations with founders, investors and operators.
我最近刚从美国出差回来，在那里，人工智能主导了我与创始人、投资者和运营者之间的大部分对话。那里的讨论已经超越了“副驾驶”（copilot）和裁员的范畴，转向了一个更有趣的问题：人工智能能否从根本上改变一家公司的运营模式？我认为可以。而且我相信，一些最大的机遇存在于那些技术历来难以渗透的行业：实物商品、碎片化的供应链、信贷和物流。
The conversation there is moving past copilots and headcount reduction toward a more interesting question. Can AI fundamentally change the operating model of a company? I think it can. And I believe some of the biggest opportunities are in the industries where technology has historically struggled most: physical goods, fragmented supply chains, credit and logistics.
我们首先构建了基础。当我们创办 BRKZ 时，我们面临的首要问题更为基础：该行业需要系统。建筑材料的采购过程极其依赖人工且高度碎片化。询价（RFQ）、定价、供应商关系、订单、交付、信贷和收款等环节散落在电话、WhatsApp 消息、电子表格、PDF 文件以及人们的脑海中。
We built the foundations first When we started BRKZ, our first problem was more basic: the industry needed systems. Building-material procurement was deeply manual and fragmented. Requests for quotation (RFQs), pricing, supplier relationships, orders, deliveries, credit and collections lived across phone calls, WhatsApp messages, spreadsheets, PDFs and people’s heads. So we spent our first three years doing something less fashionable but essential.
因此，我们花了最初三年的时间做了一些不那么时髦但至关重要的事情。我们将核心工作流程整合到了共享系统中：包括需求如何录入、供应商如何评估、报价如何生成、订单如何履行、交付如何对账、客户如何付款，以及每一笔交易如何记录。
We brought our core workflows onto shared systems: how demand enters, how suppliers are evaluated, how quotations are created, how orders are fulfilled, how deliveries are reconciled, how customers pay and how every transaction is recorded. That gave us something valuable: every workflow started leaving a digital trail.
这给了我们宝贵的东西：每个工作流程都开始留下数字痕迹。每一次询价、报价、供应商互动、交付和付款都变成了数据。随着时间的推移，这些痕迹积累成了数千万条结构化数据点，涵盖产品、供应商、交易和支付行为。
Every RFQ, quotation, supplier interaction, delivery and payment became data. Over time those trails accumulated into tens of millions of structured data points across products, suppliers, transactions and payment behavior. We don’t have to ask every contractor and transporter to change how they communicate.
我们不必要求每个承包商和运输商改变他们的沟通方式。询价仍然可以通过WhatsApp发来，送货单也可以是一张照片。智能体将这些输入转化为企业可以据以行动的结构化记录。你无法为一个尚未实现可观测的业务增添有意义的智能。
An RFQ can still arrive on WhatsApp and a delivery note as a photo. Agents turn those inputs into structured records the business can act on. You can’t add meaningful intelligence to a business you haven’t first made observable. Systemize → Capture → Understand → Automate → Agentize I’ve come to think the AI journey for traditional industries follows a progression: systemize, capture, understand, automate, agentize.
系统化→采集→理解→自动化→智能体化。我逐渐认为，传统行业的AI之旅遵循一个渐进过程：系统化、采集、理解、自动化、智能体化。所谓智能体化，我指的是赋予智能体在明确界限内采取行动的责任。
By agentize, I mean giving agents responsibility to act within clear limits. The first two steps are essential. Without data, AI has no proprietary context.
前两步至关重要。没有数据，AI就没有专有上下文。而没有上下文，你就是在把通用模型架在所有人都能获取的相同信息之上。
And without context you’re putting a general-purpose model on top of the same information available to everyone else. From prediction to execution Once we had enough transaction history, we started asking different questions. Could the system learn how we price?
从预测到执行。一旦我们积累了足够的交易历史，我们就开始提出不同的问题。系统能否学习我们如何定价？它能否判断哪些供应商与某个询价相关？
Could it work out which suppliers were relevant for a given RFQ? Pricing building materials is surprisingly hard. The same product can carry different prices depending on quantity, location, delivery requirements, timing and market conditions. Traditionally that knowledge sits in the heads of experienced procurement people.
给建筑材料定价出奇地困难。同一款产品可能因数量、地点、交付要求、时机和市场条件而承载不同的价格。传统上，这些知识存在于经验丰富的采购人员的头脑中。
So we built a pricing engine on our historical RFQs and transactions. Mizan, our AI procurement agent, is now live on our first product categories. It generates price recommendations almost instantly for procurement officers to review and adjust before submission.
因此，我们基于历史询价和交易构建了一个定价引擎。我们的AI采购智能体Mizan现已在我们首批产品品类上线。它几乎即时生成价格建议，供采购人员在提交前审阅和调整。我们花了数年时间收集的数据不再只是描述过去发生了什么，而是开始帮助我们决定接下来应该发生什么。
The data we’d spent years collecting stopped describing what had happened and started helping us decide what should happen next. Prediction is useful. Execution is more interesting. Nusa taught us that AI becomes far more powerful when it can act inside a system rather than simply answer questions about it.
预测固然有用，但执行更为引人入胜。Nusa教会了我们，当人工智能能够在系统内部采取行动，而不仅仅是回答有关系统的问题时，它会变得强大得多。代理可以掌控交易的哪些部分？
Which parts of the transaction can an agent own? Instead of asking “where can we add AI?”, we started asking “which parts of the transaction can an agent own?”A transaction breaks naturally into domains (sales, procurement, credit, operations and finance), and each has inputs, decisions, actions and exceptions.
我们不再问“我们可以在哪里加入人工智能？”，而是开始问“代理可以掌控交易的哪些部分？”一笔交易可以自然地拆分为多个领域（销售、采购、信贷、运营和财务），每个领域都有输入、决策、行动和异常处理。在采购领域，代理可以解读询价单（RFQ）、识别产品、预测价格、选择供应商并进行有竞争力的报价。在信贷领域，它可以支持承销、额度管理、风险敞口监控和催收，并在最关键的决策环节由人工进行监督。
In procurement, an agent can interpret an RFQ, identify products, predict prices, select suppliers and quote competitively. In credit, it can support underwriting, limits, exposure monitoring and collections, with human oversight where the decisions matter most. In operations, agents coordinate orders, validate deliveries and reconcile the physical world with our systems. The goal is an agentic layer across the transaction, with clear responsibility for the work each agent carries out. The hidden cost is coordination Companies don’t only have labor costs.
在运营领域，代理负责协调订单、验证交付，并将物理世界与我们的系统进行核对。我们的目标是在整个交易过程中建立一个代理层，明确每个代理所承担的工作职责。隐形成本在于协调。公司不仅有劳动力成本，还有协调成本。
They have coordination costs. A single customer request might touch sales, procurement, operations, logistics, finance and credit before it becomes a completed transaction. Each handoff creates latency. Someone sends a message.
一个客户请求在最终完成交易之前，可能会涉及销售、采购、运营、物流、财务和信贷等多个部门。每一次交接都会产生延迟。有人发送消息，有人请求审批，有人将信息从一个地方复制到另一个地方。随着组织规模的扩大，协调本身就变成了一项工作。
Someone asks for approval. Someone copies information from one place into another. As organizations grow, coordination itself becomes work. AI can remove enormous amounts of this invisible work, and I suspect that will ultimately matter more than automating individual tasks.
人工智能可以消除大量此类隐形工作，我认为这最终比自动化单个任务更为重要。
More output per person My strongest takeaway from the US was that the interesting question isn’t “how many people can AI replace?”
人均产出最大化。我从美国之行中得到的最深刻的体会是，有趣的问题不是“人工智能可以取代多少人？”，而是“每个人能多完成多少工作？”我希望人工智能能让一名优秀的销售人员具备以前需要整个支持团队才能完成的能力。
It’s “how much more can every person accomplish?”I want AI to turn a great salesperson into someone with capabilities that would previously have required an entire support team. Imagine every salesperson with a digital twin. It knows which customers are likely to reorder, which have quietly reduced their purchasing, which quotations didn’t convert and which products a customer should be buying but isn’t.
想象一下，每位销售人员都拥有一个数字孪生体。它知道哪些客户可能会复购，哪些客户悄悄减少了采购量，哪些报价未能转化，以及客户应该购买但尚未购买哪些产品。我该给谁打电话？为什么是现在？我应该卖什么，价格定多少？哪段关系在我察觉之前就已经恶化了？
Who should I call? Why now? What should I sell, and at what price? Which relationship is deteriorating before I’ve noticed? The salesperson still owns the relationship. An agent can handle preparation, analysis and routine coordination, and eventually take on follow-ups and reorders within agreed limits.
销售人员依然掌握着客户关系。智能体可以处理准备工作、分析和日常协调，并最终在约定的限度内负责后续跟进和复购事宜。同样的逻辑也适用于财务、采购和运营部门。随着智能体承担起日常工作，员工可以将精力集中在那些最能体现其判断力和人际关系价值的地方。
The same logic applies to finance, procurement and operations. As agents take on routine work, people focus on where their judgment and relationships create the most value. The physical world has to be made visible In physical industries, much of what matters still happens outside the data a company captures. A truck arriving at a site is data.
物理世界必须实现可视化。在实体行业中，许多重要事项仍然发生在公司所捕获的数据之外。卡车抵达现场是数据，托盘被卸下是数据。材料质量检测不合格、库存积压在仓库、供应商屡次迟到：这些都是数据。但如果没人记录这些事件，它们在数字层面就不存在。人工智能只能对它所能“看到”的事物进行推理。
A pallet being unloaded is data. A material failing a quality test, inventory sitting in a warehouse, a supplier repeatedly arriving late: all data. But if nobody captures those events, they don’t exist digitally. AI can only reason about what it can see. This leads to a counterintuitive idea: operational complexity can become a moat. But if one operations employee can oversee dramatically more transactions because agents handle routine coordination, and one procurement person can manage dramatically more spend, then the economics change.
这引出了一个反直觉的观点：运营复杂性可以成为一道护城河。但如果一名运营员工因为智能体处理了日常协调工作而能够监管更多的交易，一名采购人员能够管理更多的支出，那么经济效益就会发生改变。
The harder a workflow is to execute manually, the more valuable it becomes once you can teach machines to execute meaningful parts of it reliably. Don’t rebuild everything just because you can Not every problem needs the smartest model in the world.
一个工作流程手动执行起来越困难，一旦你能教会机器可靠地执行其中有意义的部分，它就变得越有价值。
The enterprise AI stack won’t have one model.
不要仅仅因为你能做到就重建一切。并非每个问题都需要世界上最聪明的模型。企业人工智能技术栈不会只有一个模型。它将拥有多个模型，而架构将决定哪种模型——本地模型、专用模型还是前沿模型——适合每一项工作。
It will have many, and the architecture will decide which one, local, specialized or frontier, fits each job. And AI makes building software dramatically easier. But if someone has already solved a generic problem extremely well, buying still makes more sense than building. Our engineering capacity goes where BRKZ has something proprietary: our workflows, transaction data, pricing knowledge, supplier network and payment history.
人工智能极大地降低了构建软件的难度。但如果某个通用问题已经有了非常成熟的解决方案，那么“购买”依然比“自研”更明智。我们的工程能力将集中在 BRKZ 拥有专有优势的领域：我们的工作流程、交易数据、定价知识、供应商网络以及支付历史。
The competitive advantage is less about which model you use and more about what context you can give it and what actions you allow it to take. AI doesn’t make strategy less important. It makes choosing what not to build more important.
竞争优势不在于你使用了哪种模型，而在于你能为其提供什么样的背景信息，以及你允许它采取什么样的行动。人工智能并没有降低战略的重要性，反而让“选择不构建什么”变得更加重要。
Where this goes The bigger opportunity is connecting the whole transaction across sales, procurement, credit, operations and finance. A customer shouldn’t need to understand BRKZ’s organizational structure.
未来趋势：更大的机遇在于将销售、采购、信贷、运营和财务等环节的整个交易过程连接起来。客户无需了解 BRKZ 的组织架构。他们只需说：“我下周二需要在该项目现场用到这些材料”，而该请求背后的所有工作都应由各个智能体（agents）协同完成。
They should be able to say, “I need these materials at this project next Tuesday,” and everything behind that request should be coordinated across agents. Push that five years out. A contractor’s procurement agent knows what the project needs and talks to BRKZ’s agent.
展望五年后：承包商的采购代理了解项目需求，并与 BRKZ 的代理进行沟通。我们的代理理解需求、客户历史、定价和供应商库存情况，并与供应商的代理进行对接。物流代理负责协调配送。财务系统达成条款。人类负责制定政策、管理关系并处理异常情况。机器则负责执行交易。
Our agent understands the requirement, the customer’s history, pricing and supplier availability, and talks to supplier agents. A logistics agent coordinates delivery. Financial systems agree terms. Humans set policy, manage relationships and handle exceptions. Machines execute the transaction. Agents will transact with agents.
智能体将与智能体进行交易。今天为这一世界构建基础设施的公司，与那些仅仅在现有软件中添加聊天机器人的公司，将会有本质的区别。技术不再是商业模式的附属品，它正在成为商业模式的操作系统。
The companies building infrastructure for that world today will look very different from companies adding chatbots to their existing software. The technology isn’t sitting beside the business model. It’s becoming the operating system of the business model.
考验在于经济效益。在接下来的几年里，每家公司都会自称是“人工智能驱动”的，因此这个词很快就会变得像说“我们公司使用互联网”一样毫无意义。我们能否在不按比例增加员工人数的情况下处理更多的交易？每位销售人员能否产出更多成果？我们能否更快地报价、做出更好的信贷决策、加快营运资金周转，并让客户以更少的阻力进行交易？
The test is in the economics Over the next few years every company will call itself AI-powered, so the term will soon mean about as much as saying your company uses the internet. Can we process significantly more transactions without growing headcount at the same rate? Can each salesperson generate more output? Can we quote faster, make better credit decisions, turn working capital faster and let customers transact with less friction? Those are the AI metrics I care about, not tokens consumed or copilots deployed.
这些才是我关心的AI指标，而不是消耗了多少Token或部署了多少Copilot。我不断思考的问题是：得益于这项技术，BRKZ能从每一位同事、每一美元营运资金和每一笔交易中获得多少额外的产出？
The question I keep coming back to is this. How much more output can BRKZ generate from every colleague, every dollar of working capital and every transaction because of this technology?
对于我们地区的机遇，我认为这对沙特阿拉伯及整个中东和北非地区的创始人尤为重要。我们拥有许多仍处于数字化早期阶段的庞大行业：建筑、制造、物流、医疗保健和批发贸易。许多人看到这一点，认为这是技术劣势。而我看到的恰恰相反。
The opportunity for our region I think this matters particularly for founders in Saudi Arabia and the wider MENA region. We have enormous industries still early in their digitization: construction, manufacturing, logistics, healthcare and wholesale trade. Many people look at that and see a technology disadvantage. I see the opposite.
我们有机会跨越整整一代软件。我们不需要花未来二十年的时间去复制在其他地方已经建立的每一个系统和工作流程。我们可以现在就将这些行业系统化，前提是假设智能和智能体从第一天起就内置于这些系统中。
We have a chance to skip an entire generation of software. We don’t need to spend the next twenty years reproducing every system and workflow that was built elsewhere. We can systemize these industries now, on the assumption that intelligence and agents will sit inside those systems from day one.
如果一个行业目前仍依赖WhatsApp、Excel、电话和人脑来运作，那么首先要让工作流程变得可观测。掌握你业务的专有背景。然后思考智能可以在哪些环节改善决策，一旦它变得可靠，再思考它可以在哪些环节从辅助人类转变为直接执行工作。
If an industry is still running on WhatsApp, Excel, phone calls and people’s heads, start by making the workflow observable. Own the proprietary context of your business. Then ask where intelligence can improve decisions and, once it’s reliable, where it can move from advising people to doing the work.
这对我们的生态系统来说是一个非凡的机遇。但这需要创始人分享更多真正有效和无效的经验：架构、错误、人类在哪些方面仍优于机器、智能体在哪些方面会失败，以及经济效益在哪些方面可行或不可行。如果我们能互相学习，整个生态系统将会发展得更快。
That’s an extraordinary opportunity for our ecosystem. But it requires founders to share more of what actually works, and what doesn’t: the architecture, the mistakes, where humans still outperform machines, where agents fail, and where the economics do and don’t work. The ecosystem will move faster if we learn from each other.
从资源出发，而非从功能出发。BRKZ 在这段旅程中仍处于早期阶段。我们正在进行实验、学习，也犯了不少错误。但有一个原则对我来说已经变得清晰。
Start from the resource, not the feature BRKZ is still early in this journey. We’re experimenting, learning and getting plenty wrong. But one principle has become clear to me. Don’t start by asking how to add AI to your company.
不要从询问如何将人工智能引入公司开始。而应从询问如果智能机器仅仅是你可用的另一种资源（与人力、资本和软件并列）时，你会如何重构公司开始。然后进行反向推导。
Start by asking how you’d build your company differently if intelligent machines were simply another resource available to you, alongside people, capital and software. Then work backwards. For us, the path has been: systemize, capture, understand, automate, agentize.
对我们而言，路径是：系统化、捕获、理解、自动化、代理化。对于其他公司，路径可能有所不同。但目标应当是一致的。不要仅仅利用人工智能让旧的工作方式效率稍有提升，而要利用它去发现一种从根本上更好的运营方式。
For another company it may look different. But the objective should be the same. Don’t use AI to make the old way of working slightly more efficient. Use it to discover a fundamentally better way of operating.
如果该地区的更多创始人能以这种方式思考，并公开分享他们的经验，其影响将远远超出任何一家公司。这正是我希望本文能促成的对话。
If more founders in our region approach it that way, and openly share what they learn, the impact will extend far beyond any one company. That’s the conversation I hope this contributes to.
投稿文章。非 TNW 新闻编辑部制作，不代表 TNW 的编辑立场。
Contributed article. Not produced by the TNW newsroom and does not reflect the editorial stance of TNW.