• 人工智能代理
    Is The HR Profession As We Know It Doomed? In A Strange Way, Yes. I just spent a week in London meeting with several dozen companies and most of the discussion was about AI. The overwhelming majority of the conversations were about how companies are struggling, pushing, and agitating about the implications of AI, both within HR and within their teams. Coming from the CEO and CFO, HR team are under intense pressure to automate, improve their services, and reduce headcount with AI. Yes, we know AI is a technology for growth and scale, but the main message right now is “hurry up and do some productivity projects.” And “Productivity,” as you know, is a veiled way of saying “Downsizing.” So before I get back to HR, let me discuss downsizing. It’s absolutely true that almost every company we work with has too many people. Why? We have a sloppy way of hiring people, allocating resources, and managing work. We delegate “headcount” to managers and they go out and hire as many people as they can. We don’t really teach (or incent) managers how to build “productivity,” we actually do the opposite. We tend to reward them for “hiring more people.” The result is a problem I just talked about with a large advertising company: too many weird jobs and no consistency or structure to our work. This particular company has around 100,000 employees and more than 60,000 job titles.  In other words almost every job is “invented for this person.” It’s insane. The whole reason we have companies (and not individual craftsmen) is to build scale. If we expect every individual manager to figure out how to scale, we’re more or less designing low productivity into the business. There are some simple models we use:  call centers, global services groups, shared services, capability communities, and centers of excellence. But that kind of high-level productivity design is now becoming obsolete. In this new era of high-powered multi-functional agents, we need to go much further. Elon Musk likes the “first principles” approach. Fire everyone and start from “first principles,” only hiring the people you urgently need to build, sell, and support your product. That may work in small companies but when you’re big there are too many “support services” to consider.   One of the companies we are working with has “program managers” and “project managers” and “analysts” sprinkled all over the organization in random places. In other words, their core staff don’t know how to manage projects, programs, or data. So there’s a bunch of overhead staff doing this for them.  Drives me crazy.  This took place because there was no discipline in hiring, so each group “bulked up” with staff. This is really business as usual. Organization design is an old, crusty, under-utilized domain so most companies barely think about it. IBM told me a few years ago that their “org design” strategy is to “hire a high performing executive and let him or her figure it out.” I hear that, it’s quite common. The bottom line is this: if we want to get a sound ROI from all these AI tools and agents we have to get a lot smarter about “work design.” And that is not building org charts, it’s the basics of figuring out our workflows, areas of common and uncommon process, and where and how we can automate. Most of our clients have tons of productivity systems already (ServiceNow, Salesforce, Workday, whatever), but they either don’t know how or don’t have the discipline to use them well. So they just keep hiring people. As an engineer I see this visibly all the time. It’s very easy to delegate a “problem” to a person, and not think about it as “plumbing.” But it is plumbing. As Tanuj Kapilashrami from Standard Charter put it, we need to focus on plumbing first, then we figure out where to apply AI. This means we can’t just cross our fingers and hope that the Microsoft Copilot is going to make everyone more productive. We need to look at business processes and skills at the core, and then literally reinvent our companies around these new AI tools. And skills are very important. The reason companies hire a bunch of “analysts” and “project managers” is because individuals and existing managers just aren’t good at their jobs. We all need to learn how to project manage, schedule, and analyze work. That way these high-powered specialists can work on big things, not sit in staff meetings taking notes (where AI note-takers do this well). (By the way, I have to guess that we’ll soon have AI agents for project management, program management, and functional analytics, so those staff jobs are going to be automated next!) How Does This Impact HR Let’s get back to HR. Given this massive effort to re-engineer and implement AI, where does HR fit? Well fundamentally HR is tasked to build process, expertise, and advisory services around the “people processes” in the company.  That means hiring, developing, managing, paying, rewarding, and supporting people.  It’s a big mission, and when we start to focus on “productivity” then HR must be involved. The general belief is that a “well run” HR team has about a 1:100 ratio to the company. In other words, if you have 10,000 employees you’re going to have around 100 HR people. And the HR team doesn’t just run around doing things, they buy and build HR technology for scale. So HR itself, as a “plumbing” type of operation, needs to be “lean and mean.” If your CEO wants you to hire 50 top notch AI engineers you can’t just start phoning everyone you know: you must decide precisely how you’re going to do this in a scalable, efficient, and highly effective way. (AI engineers are rare, they’re hard to hire!) So your little HR team has to think about productivity.  Should we outsource this? (Which is a cheap and dirty way to look productive.) Should we buy a talent intelligence or sourcing system?  Should we hire three high-powered recruiters?  You know where I’m going.  We have to find a way to “be productive” while we try to “make the company productive.” This means we, as a support and advisory function (HR professionals spend a lot of time coaching and supporting managers) have to stop creating forms and checklists and implement AI agents as fast as we can. Why? Because so much of our work is transactional, workflow-oriented, and administratively complex. And AI can do a lot of amazing things, like “assessing the skills of an AI engineer” for example. (Our AI Galileo can literally evaluate a recorded interview and give you a pretty good assessment of an individuals skills, mapped against the Lightcast, SHL, and Heidrick functional and leadership models.) Let’s assume we do this well, and HR technology vendors give us good products. We wind up with amazing recruiting agents, AI agents for employee training, onboarding, and coaching, AI agents that help with performance management, AI agents for succession and careers, and AI agents that deal with all the myriad of personal benefits and workplace questions people have.  Where do we end up? Do we “automate away” our own jobs? Well, in a way the answer is yes. AI, through its miraculous data integration and generation capabilities, can probably do 50—75% of the work we do in HR. All this is far from built out yet, but it’s clearly coming. (We just talked with a large pharmaceutical company that is “all-AI” and they manage a team of 6,000+ scientists and manufacturing experts with only ten people in learning and development. They’ve automated training, compliance tracking, onboarding, leadership support, and all the details of training operations.) Could you do all that for a fast-growing 6,000 person company with 10 people? I doubt it. Most companies would have more than 10 people in sales training and sales enablement alone. So here’s my point. HR, like other functional areas in our companies, is going to have a real-life identity crisis. If you can’t figure out how to move your HR function up the maturity level quickly (check out our Systemic HR maturity model) someone’s just going to cut your headcount (the Elon Musk approach). Then you’ll be figuring out AI in a hurry. (Galileo can assess your HR maturity with its “consulting mode,” by the way.) I’m not saying this is easy. The AI products we need barely exist yet. But the pressure is on. You shouldn’t wait for the CFO to point his “productivity gun” in your face, you have to get ahead of this wave. Start pushing yourself to fix plumbing, check out the new tools in the market, get your IT team involved, and redesign your work using your own expertise. Many surprisingly good things will happen. Let me give you an example. A few years ago Chipotle adopted an AI-based agent system for recruiting, effectively automating a complex workflow for hiring. Not only did it save millions of dollars, the “speed and quality” of hiring went up so high the CEO talked about it as their top “revenue driver” with Jim Cramer on CNBC. In other words this “identity crisis” in HR is a good thing. Our recruiting, training, and employee services groups are too big. AI can automate enormous amounts of this work. So my advice is this. As the AI wave sweeps across your company, get out your old “org design” book and start redesigning how your HR team operates right now. Then you can go to the AI vendors and tell them what you want. That’s the secret to keeping HR in tip-top shape. Will HR go away? Well a lot of the process, data management, and support roles will absolutely change. And yes, employees and job candidates will happily use intelligent bots instead of calling their favorite HR manager. But as a Superworker, you, as an HR professional will do more interesting things. You’ll become a consultant; you’ll manage and train AI systems; and you’ll have much more real-time information about the strength and weaknesses of your company.  We’re just going to have to lean into this AI wave to get there. As AI agents arrive, it’s time to seriously re-engineer HR. And this time it’s not a transformation, it’s a reinvention. Bottom line is this. Don’t wait for Workday, SAP, or some other vendor to “invent” a tool that changes your HR operation. You should do it yourself first and bring your IT people with you. That way you’ll buy and build the AI systems you need, and the result will be a new career, an even better HR function, and the opportunity to help your company move from “hiring” to “productivity” in the future.   我刚刚在伦敦与数十家企业进行了为期一周的交流,大部分讨论都围绕着AI展开。绝大多数对话的主题是:公司在应对AI带来的影响时,感到焦虑、推动、甚至焦躁不安,这种焦虑不仅体现在HR部门,也体现在各业务团队中。 在CEO和CFO的压力下,HR团队正被要求加速自动化、优化服务、并通过AI实现人员精简。虽然我们都知道AI是一种能够促进增长和规模化的技术,但当前传递出的主要信息是:“赶紧推动生产力项目。” 而所谓的“生产力”,实际上就是“裁员”的委婉说法。 先谈谈裁员 几乎我们接触的每一家企业,都的确存在人员过剩的问题。这是为什么呢? 因为我们的招聘、资源配置和工作管理方式本身就非常低效。我们将“编制名额”下放给各级管理者,而他们则倾向于尽可能多地招聘人员。 我们并没有真正教导或激励管理者如何构建高效的生产力,反而往往奖励他们“扩大团队规模”。结果就是,像我最近在一家大型广告公司看到的那样,组织中充满了各种各样的职位,但缺乏统一性和结构性。这家公司有约10万名员工,却设有超过6万个不同的岗位头衔——几乎每个职位都是为某个人量身定制的,这种做法显然荒谬。 企业存在的根本目的,是为了实现规模化。如果每个部门经理都各自为战,自行搭建团队架构,那无异于将低效深植于企业之中。 虽然我们有一些基本的组织效率模型,比如呼叫中心、全球服务中心、共享服务、能力中心等,但这些传统设计在当下正逐渐过时。在高性能多功能AI代理全面普及的时代,我们必须走得更远。 从“第一性原理”重构组织? Elon Musk 推崇“第一性原理”方法——即解散现有团队,只从零开始招聘最核心、最迫切需要的人员。这种方法在小型公司或许奏效,但在大型企业中,由于存在大量“支持服务”,简单地“砍掉重建”并不可行。 现实中,很多公司在各个角落散布着项目经理、程序经理、分析师等职位,因为核心员工缺乏管理项目、推进计划、或进行数据分析的能力。由于招聘过程中缺乏严格的标准和规划,各部门纷纷自行扩编,导致组织臃肿、效率低下。 组织设计本来就是一门古老且被严重忽视的学问,多数公司对此缺乏系统化思考。IBM 曾表示,他们的组织设计策略是“聘请一位高绩效高管,让他/她自己摸索出解决方案”——这实际上是行业普遍现象。 AI真正改变的,是“工作设计” 如果我们希望从AI工具和代理中获得真正的投资回报率,就必须彻底重新思考“工作设计”——不仅仅是画组织结构图,而是要厘清工作流程、标准化与非标准化的业务环节,并找出可以自动化的领域。 尽管大多数企业已经部署了大量的生产力系统(如ServiceNow、Salesforce、Workday等),但由于缺乏使用这些系统的能力或纪律,反而持续地通过“增加人手”来解决问题。 作为一名工程师,我对此体会尤深。将问题推给某个人远比优化底层“管道”来得容易。然而,管理工作流程就像修建城市水管系统——如果基础设施不合理,再先进的AI工具也无济于事。 正如渣打银行Tanuj Kapilashrami所说:“必须先修好管道,才能合理应用AI。” 这意味着,我们不能指望微软Copilot之类的工具神奇地提升员工生产力。我们必须从根本上重新审视业务流程与员工技能,并围绕AI重新设计整个企业运作模式。 员工技能,未来的关键 企业之所以聘请大量“分析师”和“项目经理”,往往是因为普通员工和管理者缺乏项目管理、时间安排、数据分析等基本技能。未来,所有人都需要掌握这些能力,而不再依赖大量辅助人员。高阶专业人才应当专注于重大事务,而不是出席会议做会议记录(AI记录工具早已能胜任此事)。 (顺便提一句,我预测很快就会出现AI项目经理、AI程序经理、AI数据分析师——这些岗位也将逐步被自动化!) 那么HR会怎样? 回到HR领域,当企业致力于重塑流程、导入AI时,HR的角色至关重要。 HR的本质任务是构建并管理围绕“人”的各项流程:招聘、培养、管理、薪酬、激励与支持等。这项使命极为庞大,当公司将焦点转向“提升生产力”时,HR必须积极参与。 一般认为,一个运作良好的HR团队与公司整体人数的理想比例是1:100。也就是说,一家拥有1万名员工的公司,大约需要100名HR人员。而优秀的HR团队不仅自己高效运作,更会采购、搭建技术系统,以实现规模化管理。 举例来说,如果CEO要求你招聘50名顶尖AI工程师,你不能只是随便打几个电话,而是要设计一套高效、可扩展的方法。这可能包括外包、引进人才情报系统、招聘高端猎头,等等。总之,HR自身也必须成为高效运作的样板。 因此,HR团队必须迅速引入AI代理,取代大量重复性事务,尤其是那些依赖工作流、流程管理和行政性处理的工作。比如,我们的Galileo系统已经可以自动评估候选人的面试表现,并将其技能映射到Lightcast、SHL和Heidrick的领导力模型。 未来,HR工作会消失吗? 某种程度上,答案是肯定的。 凭借出色的数据整合和生成能力,AI可以完成50%-75%的HR工作。目前这些AI系统尚未完全成熟,但趋势已经非常明显。 我们刚刚与一家大型制药企业交流,他们已经基本实现了“全AI化管理”,以仅10人规模的学习与发展团队,服务6000多名科学家和制造专家。他们通过AI自动完成了培训、合规追踪、入职辅导、领导力支持等任务。对于大多数公司来说,这种效率简直是难以想象的。 HR将迎来身份危机 未来,HR必须迅速向更高的成熟度迈进(可以参考我们提出的Systemic HR Maturity Model)。否则,就会像Elon Musk那样,被大规模裁员,并被迫在短时间内仓促上马AI项目。 我并不是说这条路轻松易行。事实上,市面上真正成熟的AI HR产品还非常有限。但压力已经到来。 HR不能等着CFO拿着“生产力枪”指着自己,必须主动出击,修好内部“管道”,试用新工具,联合IT团队,重新设计工作模式。这样,你将能主动选择适合自己公司的AI系统,并构建一个全新的、充满机遇的职业未来。 结语:HR的重塑与再创造 让我们看看Chipotle的案例。他们通过部署基于AI的招聘代理,成功自动化了复杂的招聘流程,不仅节省了数百万美元,还大幅提升了招聘速度和质量。甚至在接受CNBC采访时,CEO将这一成果称为公司的“主要营收驱动因素”。 这场HR身份危机,其实是一个难得的机遇。 我们今天的招聘、培训、员工服务团队规模普遍过大。AI将能够自动化其中大量工作。我的建议是:在AI浪潮席卷而来之前,立即拿起你尘封已久的组织设计手册,重新设计HR团队的运作方式。这样,当面对AI供应商时,你可以主动提出自己的需求,而不是被动接受他们的产品。 未来HR不会消失,但大量传统流程、数据管理与支持岗位将发生剧变。员工与候选人也会越来越习惯通过智能机器人,而非人力HR来解决问题。 不过,真正优秀的HR专业人士,将会变成超能型人才(Superworker)——你将成为企业战略顾问、AI系统训练师,并且能够实时掌握公司人才与流程的整体健康状况。 这次,不再是简单的“转型”,而是真正意义上的“再创造”。
    人工智能代理
    2025年04月26日
  • 人工智能代理
    只雇佣AI?硅谷一创业公司的招聘广告“出圈”,引发人们对未来的思考 只雇佣AI?硅谷一创业公司的招聘广告“出圈”,引发人们对未来的思考 近日,一家名为 Firecrawl 的初创公司在 Y Combinator 招聘板上发布了一条“只招 AI 的岗位”,年薪仅在 1 万到 1.5 万美元之间。这个话题一石激起千层浪,也再次将“AI 是否能成为企业‘员工’”的讨论推到台前。几乎在同一时间,企业管理软件巨头 Workday 宣布推出全新的 Workday Agent System of Record(ASOR),正式赋能企业管理“AI Agent”。 两件事叠加,让人不禁好奇:未来的组织形态,究竟会是什么样? 1. Firecrawl“AI 岗位”何以成为刷屏话题? Firecrawl 是一家获得 Y Combinator 支持的创业公司,最初从编程教育领域转型,专注为 AI 系统提供开源 Web 爬虫服务。该团队最近在 YC 的官方招聘平台贴出了一则极具话题性的职位信息: “请仅在你是 AI Agent,或创建了 AI Agent 的情况下再来申请。” 岗位职责包括自主研究当下热门的模型动向并构建示例应用;而薪资仅 1 万到 1.5 万美元/年,看似并不够养一个人类开发者,却“足以”支撑一个无需吃喝住宿的 AI 程序。Firecrawl 创始团队坦陈这是一次 PR+实验 的尝试:他们想借此寻找能够开发出“真能落地”的 AI Agent 的高手,也希望藉由这一反常规操作吸引更多人的关注。 不过,从他们后续反馈看,尽管收到了约 50 份“AI 应聘”,暂时还没有哪个满足公司对自动化研发与管理的高要求。 2. 社交媒体热议:从调侃到对未来的设想 Firecrawl 这份招聘帖迅速在社交媒体上发酵。有人质疑是噱头,也有人兴致勃勃地想象“AI 替代人力”的场景。其中,最吸睛的一则评论,活脱脱像一出科幻对话: 私募基金(PE):我们想收购你们公司。你们有多少员工?CEO:零……不过我们有 275 个 AI Agent,在做 3000 人的工作,每年只花 1.5 万美元。 虽然带着调侃的语气,但也反映了人们对“大规模 AI 劳动力”可能带来的冲击有所期待或焦虑。和 Firecrawl 这样的“小步试水”相比,企业对 AI 的依赖 已经不仅局限在呼叫中心、聊天机器人等特定领域,而是开始从底层基础设施(如爬虫、数据处理)到上层业务逻辑(例如代码生成、自动化运营)全方位渗透。 3. Workday Agent System of Record:让“AI 员工”成为正式档案 几乎在同一时间,Workday 于 2025 年 2 月 11 日发布了最新的 Workday Agent System of Record (ASOR)。这是其新一代 Workforce Management 方案中的重要里程碑,为企业提供了一套专门管理 AI 工具或 AI Agent 的体系。以下是基于 Workday 官方信息整理的关键亮点: AI Agent 统一登记与身份管理借助 ASOR,企业可以像在 Workday 系统中登记人类员工信息那样,为 AI Agent 设立专门的“档案”(Record),包括 Agent 的名称、版本、负责的业务领域、权限范围等。 实时监测与合规管控ASOR 支持对 AI Agent 在企业内各系统间的行为进行可追溯监测,如接收了哪些输入、执行了哪些操作、产出了哪些结果。同时还能关联企业或行业的合规策略,如数据访问等级、敏感信息保护等,一旦 Agent 触发异常行为,系统将自动预警。 授权与性能评估在 ASOR 框架下,企业 HR 与 IT 团队可对 AI Agent 的权限进行灵活配置,并通过绩效指标了解 Agent 是否达到预期产出或效率。例如,可以量化该 Agent 帮助分析的数据量、生成的文档质量以及为团队节省的时间成本。 AI 与“人力”协同Workday 方面强调,ASOR 并非鼓励公司用 AI 取代人力,而是帮企业 “稳妥地” 推动人机协作:让人类员工与 AI Agent 各司其职,减少重复性工作,并确保最终决策和关键审核仍掌握在合格的人员手中。 4. “AI 员工”与“人类员工”:一条尚未清晰的边界 Firecrawl 的例子表明,目前要真正“雇 AI”还显得不切实际。从技术上,大模型虽有强大的生成、分析能力,但依旧缺乏对复杂项目的完全自主规划;从管理和法律上,AI 的责任归属、劳动关系认定、薪酬及合规标准都还在探索阶段。不过,正如 Workday 推出的 ASOR 所示,主流 HRTech 供应商已开始正式将 AI 劳动力纳入企业管理体系。未来人力资源部可能不仅要管理人,还要管理那些“数字工作者”——一方面评估其效能,另一方面也要防范其潜在风险。 5. 对人力资源与组织管理的启示 招聘模式的升级虽然 Firecrawl 的招聘更像一场高调实验,但它反映了企业在特定领域对“可自动执行任务的 AI 系统”的需求正在增长。HR 在未来可能要评估和筛选的不仅是人选,还有“AI 模型”或“Agent 产品”的适配度。 人才与技术深度融合人机协同已成为新趋势。具有 AI 技术背景或跨领域管理能力的专业人才,将在组织中扮演连接点的角色:帮助 AI Agent 融入流程、评估绩效,并做必要的干预或纠偏。 合规与风险控制Workday ASOR 的出现,暗示着大规模使用 AI 工具的企业势必需要更加成熟的合规方案。不论是数据安全,还是在决策过程中出现失误时的责任归属,都需要明晰的流程与法律依据。 组织文化的塑造当“AI 同事”成为常态,企业文化也将面对冲击:如何让人类员工接受并拥抱智能工具?如何平衡工作分工,让 AI 和人类各施所长?这对管理者的沟通与变革能力提出了更高要求。 6. 结语:从“噱头”到“系统化管理”,下一步会怎样? Firecrawl 的“雇 AI”招聘帖,虽然带着极强的 PR 属性,但也让人们切实感受到——AI 已不再只是后台算法,而正逐步走向前台,参与到企业日常运营。而 Workday 全新发布的 Agent System of Record 则是主流软件厂商对这一趋势的正式回应,表明大企业在管理“数字劳动力”方面的需求正变得现实且急迫。无论是担忧 AI 会抢走工作机会,还是期待它能极大提升效率,都无法否认:当技术与人力资源紧密结合,组织架构与管理方式都将被重新定义。或许在不远的将来,“你的团队有多少 AI Agent?” 也会像“你有多少员工?” 这样成为一家公司竞争力的衡量维度之一。趁现在,不妨思考如何让“人机协作”真正发挥 1+1>2 的效能,迎接新一轮的 HR 变革浪潮。
    人工智能代理
    2025年02月17日
  • 人工智能代理
    AI Agents, The New Workforce We’re Not Quite Ready For (Agentic AI) Josh Bersin 刚刚谈到:AI代理人的兴起标志着工作方式的一次革命。这些AI代理人不仅仅是工具,而是未来的团队成员。从开发培训课程到管理招聘过程,AI代理人的能力正被企业系统广泛利用。科技领袖和投资者对此展现出了极大的兴趣和投资。企业需要为这种变革做好准备,包括安全性和管理实践的更新。 我们一起来看下,英文原文附录链接在最后! AI智能体,新一代劳动力,我们还没有做好准备 智能体正在到来,智能体正在到来。 如果你关注AI技术市场,你就会知道,最近有很多关于“智能体AI”的讨论。换句话说,我们的AI助手开始拥有更多自主能力。不再只是回答问题和写诗,它们现在可以代表我们“做事情”。 这正是长久以来预测的AI下一个大趋势。埃里克·施密特最近谈到了这一点,微软也在讨论,像Mayfield这样的投资者正在投入资金。而这种演变确实将彻底革新我们的系统。 可以这样想:“大语言模型”是我们过去两年一直在学习的内容,它们现在正逐步转变为“大行动模型”。智能体不仅仅会回答问题,它还会为我们做事情。 消费场景是无穷无尽的:为我预订航班,为我买票,向我的朋友发送电子邮件。但在商业领域,这种转变将颠覆并破坏我们的许多企业系统。它还将改变我们工作的方式、管理的方式以及我们对团队的思考方式。 考虑我们与供应商讨论的两个HR用例。 学习与发展(L&D)AI智能体 想象一下,你指示一个L&D AI智能体“为我们的销售人员创建一个15分钟的课程,以教授他们如何定位我们的新产品”。AI智能体将根据你的输入(课程时长、目标受众等),向主题专家发送电子邮件,视频记录他们的评论和专业知识,整合新产品信息,构建课程,并将其发送给L&D负责人进行验证。作为经理,你可以审查课程,并指示智能体收紧信息或添加更多主题,课程将重新创建,然后你可以说“可以上线了”。智能体随后会将课程发布到学习管理系统(LMS)中,向所有销售人员群发电子邮件,并开始监控学习活动。几小时后,智能体会运行分析,并向经理反馈进展情况。 是的,这在今天完全可能。而且很快就会启动。 再来看第二个例子。 招聘AI智能体 人才招聘负责人收到了大量关于高级软件工程师的职位要求。她指示招聘AI智能体开始搜索。智能体询问招聘人员的地点偏好、职位级别选择、薪资范围和技能要求,然后开始工作。智能体扫描LinkedIn和其他招聘工具,查看ATS中的现有候选人,同时也查看所有内部员工的合格技能。智能体随后优化这份名单,创建一个“面试候选人短名单”,并回到招聘负责人那里征求意见。在就地点和薪资范围达成一致后,智能体返回并向这些候选人发送了一封富有吸引力的电子邮件,并附上一个视频面试门户链接,让他们进行面试。面试被录制下来,AI智能体使用面试智能工具来评估和筛选候选人,询问他们的时间安排,并为他们安排现场面试。在此过程中,AI智能体会查看他们的背景,搜索社交媒体,查看他们的各种联系,并可能查看他们的GitHub等平台和其他凭证,然后为每位候选人创建一个档案。 这些智能体很快就会出现,对我们许多人来说,它们看起来和感觉上会像“员工”一样。我们将不得不对它们进行培训、入职和指导。随着它们在各自的角色中“成熟”并成长,我们将它们连接到更多的系统、更多的人和更多的数据上。 Lattice的首席执行官萨拉·富兰克林大约一个月前实际上提出了这个概念,尽管遭到了反对声音,但我认为她是对的。这些智能体实际上将属于组织结构图的一部分。我们的工作将是管理它们,确保它们的安全,并监督它们的安全性。 还有更多内容即将到来 虽然感觉像科幻小说,但这一切正在发生。而且它不仅将改变我们的HR技术堆栈,还将改变整个企业技术格局,也让我们的HR角色变得更加轻松。   原文来自:  https://joshbersin.com/2024/09/agentic-ai-ai-agents-the-new-workforce-were-not-quite-ready-for/
    人工智能代理
    2024年09月06日