如何用AI推动业务增长:识别瓶颈、定义智能体角色与优化数据
9 min Machine translated Money & business
9 min Machine translated Money & business
人们以为关键在于工具,他们以为关键在于模型,他们以为关键在于人工智能,但实际上,关键在于你如何解读这些信息以及如何确定优先顺序,这才会成为改变局面的关键决定它对你是否有用。
外面有很多人在使用人工智能,但用错了方法,因此他们不会成长,也赚不到钱。
我首先要和你分享的信念与理念之一是这样的:五条中的第一条,就是大多数人用人工智能来更快地干蠢事。
而这不是我们取胜的方式。我们不会更快地做错事,我们会做正确的事。
要做到这一点,你得明白什么是制约因素。任何企业在任何时候,都只有一个主要制约因素。
如果你用人工智能找出它,这样你就能让整个团队全力以赴,一起解决那个问题。
这种思维方式的转变,是因为我教大多数人一些强大的 AI 技术之后,他们只是一味地把错误的事做得更快,业务没有增长,毛利率也没有提高。
第二个重要的思维转变是你加上 AI,胜过你的团队加上 AI。
我知道你希望团队成长,但团队对 AI 的采用,是你先采用之后的结果。
尽管我指导过许多杰出的企业主,他们的营收达到一千万、两千五百万、五千万,我教他们用更先进的方式运用 AI,来获取大量需求,并自动化销售、流程和交付,从而创造利润,他们还是会先想到自己的团队。
当我告诉他们: “你的团队正跟着你的步伐走。”
瓶颈在你,不在你的团队。所以我鼓励你在日历中留出时间,专门用于学习,用于成长,也用于成为团队需要的那个人。
这是个很重要的主题,也是个重要原则。别让它被忽视。
第三,AI思维模式的一大转变,在于角色设计,而非提示词工程。
我认为,整个AI世界都变了,当智能体开始由一些大型模型托管,比如Grok Bot、ChatGPT Dots,以及刚推出的Claude Duet,而且大多数人还无法轻易使用它。
我们也在Muse和Instinct上看到了这一点。当人们问: “嘿,Dan,我怎么才能让我的AI 做更聪明的事?”
我会说:“接下来你想雇谁,才能为你的业务争取到最多的精力和时间?”
角色。” 然后我请他们定义那个人会做什么。
当他们向我描述这个角色时,如果他们拿出手机,然后跟他们的 Grok Bot 对话,跟他们的 Claude Duet,跟他们的 ChatGPT Dots 对话,他们很快就会明白,自己做的全都只是定义这个角色,而不是编写提示词,让 AI 接下来替他们行动并执行任务。
这是一个巨大的思维转变,大多数人,外行们,永远无法取得成果,因为他们还困在这里。
他们还停留在 2026 年 2 月,那并不久远,但变化就是这么大。
所以,当你思考如何围绕 AI 思维推进部署时,请考虑到,你今天作为 CEO,作为企业主,有责任更善于定义智能体的角色。
也就是说,你得学会把角色定义清楚。你们很多人难以扩大公司的原因是你们无法明确公司需要哪些岗位,来帮助你们突破最大的瓶颈。
接下来是一个概念,我称之为数据权益。你看,在人工智能领域,权益归根结底取决于数据质量。
拥有和其他人一样的模型和工具,并不能让你占据优势。
那么,机会在哪里?机会就在于清理并核实数据,这样你就能把数据输入这些模型,让它们把一个想法一路推进,直至完成目标,全程无需人工参与。
好吧?前几天,我的朋友问我他是如何管理整个销售流程的,目前,他们甚至还没开始通话,就已经有一份登记表了,接着,他们会进行销售沟通,然后他们得签署一份,呃,协议,接着得为他们办理入职,然后得把他们录入系统。
我就说:“那你有这些步骤的示例吗?” 他说:“没有。”
我说:“问题在于,如果你没有定义好数据,也没有清理好数据和准备好示例,可以把这些提供给它,让它运用那枚执行力的热寻导弹,那你就没有价值,也就谈不上股权价值。”
每家企业都需要开始问自己: “我该如何在企业内部打造数据资产价值?”
因为这会让潜在买家觉得你更有价值。这也会让人工智能真正替你把工作完成。
最后一点,绝对不是最不重要的一点,是学习循环。
人们总是跟我说: “Dan,我有机器人。我有机器人。”
我说:“不,你没有。你用的是自动化。” 你没有机器人也没有智能代理除非你有一个学习循环。
学习循环是一种流程你在其中定义一个目标,它会生成结果,并在某个环节收到反馈。
这些反馈会被送回最初的提示中,然后它会用新知识来执行流程中的各个步骤,这样它就会不断进步,越来越好,越来越好。
所以,如果你们公司和团队里目前有任何智能代理或机器人,但团队没有就它的工作给它反馈,那你就真的没有让它随着时间不断进步的能力。
这就是最优秀的公司借助 AI 扩大规模的方式,他们明白,智能代理若非如此,就不算智能代理没有学习循环,并建立起这个流程。
一个简单的例子是,让你的智能体做某种分析,把分析发到 Slack,请团队对这份分析提供反馈,然后让智能体监控它在 Slack 发帖时的情况,收集反馈,并把它视为升级其大脑的机会,升级它的规则、步骤、流程和质量。
然后,如果你的团队在查看输出时给出反馈,而你把反馈输入智能体,每周一次,一年52次,它就能变得越来越好。
这也会影响你的数据公平性。你看出这些环节是如何层层叠加的吗?
我们知道,如今的人工智能仍然只是让大多数企业主更快地做错事或蠢事,而且我们必须提升自己的技能并加深团队的理解,团队的成长永远不能超越我们的能力,因为我们是领导者,也就是瓶盖,而且我们需要做得更好,明确各个智能体的职责,
还得理解智能体的数据层面,它们将如何使用这些数据,数据是否可靠,以及我们是否有很好的实例,我们是否知道什么样才算赢,我们是否知道什么样才算成功,然后把这些学习反馈循环融入流程,这就是我们赚大钱的方法。
这就是你获胜的方法,这已经不再是可选项。
2027年将成为分水岭,区分那些决定走向专业化的公司,和最终倒闭的业余玩家。
如今,当我们想到每家公司都将成为一家人工智能软件公司时,这不是 SaaS,不是软件即服务。
而是把你的服务变成软件。这是你为市场解决的问题,把相关信息传达得炉火纯青,然后在后台,用它以更快的速度交付,并提供更高的质量,从而提高你的毛利率,因为你不必雇一大群人。
让员工继续做需要人来做的事,用 AI 来处理电脑上的工作,这样一来你就能赚很多钱。
说快一点,你已经看到最后了。让我告诉你我对你的了解。
你不是那种只想捡现成好处的人。你不是那种只会在社交媒体上不停刷负面内容的人。
你来这里是因为你想创业,建立一家真正的企业,而且你早已受够了无法获得真正的见解来推动业务增长。
所以我想帮你免去摸索这件事所经历的所有痛苦免得像我一样走弯路。
所以我整理了我学到的一切,放进了《Scale Workbook》里。
这是我和每位合作的创业者一起使用的内部文件它能把混乱的生意变成赚钱机器。
教你这一切,不用再耗费更多时间。完全免费。
链接就在简介里,或者你可以直接扫描屏幕上的二维码因为我在这里,就是为了帮你打造自己的商业帝国。
End of transcript · 1,484 words · machine translated
People think it's the tools, they think it's the models, they think it's the AI, but it's actually the way you're interpretating the information and prioritizing that's going to make it a game changer for you or not.
There's a lot of people out there that use AI, and they use it the wrong way, and they will not grow or make money because of it.
And one of the first beliefs and philosophies that I have to share with you is that, number one out of five, is that most people use AI to do dumb shit faster.
And that is not the way we win. We don't do the wrong thing faster, we do the right thing.
And to do that, you need to understand what is a constraint.
And every business, at any point in time, only has one primary constraint.
And if you identify it using AI, then you can surge your whole team on solving that problem.
That mindset is a shift, because most people, I teach them some of these powerful AI technologies, and they just keep doing the wrong thing faster, and their business doesn't grow and their gross margins don't grow.
The second big mindset shift is you plus AI over your team plus AI.
I know you want your team to grow, but your team adoption is a byproduct of your adoption.
And even though I've coached many incredible business owners that are doing 10 million, 25 million, 50 million on the advanced ways to use AI to extract large amounts of demand and automate sales and processes and delivery so they can create profit,
they still default to their team. When I tell them, "Your team is following your lead.
You are the bottleneck, not your team." So I encourage you to create space in your calendar to dedicate to your learning, to your growth, to you becoming that person for your team.
Big theme, big principle. Don't let it go unnoticed.
Number three, big shift in the mindset of AI is its role engineering over prompt engineering.
The whole world of AI shifted, I believe, when the agents started getting hosted by some of these large models like Grok Bot, like ChatGPT Dots, like Claude Duet that just came out and is not readily available for most people.
And we're seeing this with Muse and Instinct.
And what happens when people go like, "Hey, Dan, how do I get my AI to do smarter things?"
I go, "Who would you want to hire next to give you the most bandwidth in your business?"
Role. And then I ask them to define what that person would do.
As they're defining it to me, if they pulled out their phone and they would talk to their Grok Bot, to their Claude Duet, to their ChatGPT Dots, they would learn quickly that all they're doing is defining the role, not the prompts for the AI to now do and execute on their behalf.
That is a massive mind shift that most people out there, the amateurs, will never get results because they're still stuck here.
They're still stuck in February 2026, which is not long ago, but that's how much it's changed.
So as you think about your rollout around the mindset of AI, please consider that your responsibility today as a CEO, a business owner, is to get better at role definition of your agents.
That is, you got to get good at defining it.
The reason why a lot of you guys struggle with scaling your company is you can't define the roles that would come into your business to help you unblock your number one constraints.
The next one is this concept I call data equity.
See, equity in the world of AI comes down to the quality of your data.
Having access to the same models and the same tools that everybody else has doesn't give you an advantage.
So where is the opportunity? The opportunity is in you cleaning up and verifying the data so that you can feed these models so they can take an idea all the way to a completed goal without a person involved.
Okay? When my friend the other day asked me about how he manages his whole sales process, where right now before they even get on a call, there's an intake form, then they have a sales conversation, then they have to sign a, uh, an agreement,
then they got to onboard them, then they got to put them in the system.
I'm like, "Well, do you have examples of those steps?"
And he's like, "No." I go, "Well, the problem is, is that if you don't have the data defined and it's cleaned up and you have examples you can give it so that it can use its heat-seeking missile of execution, then you don't have value, which is the equity part."
Every business needs to start asking themselves, "How do I create data equity inside my business?"
Because that's what's going to make you valuable to a potential buyer.
That's what's going to make the AI actually get the work done for you.
The last one, and definitely not the least, is learning loops.
People tell me all the time, "Dan, I got bots.
I got bots." I said, "No, you don't. You have automation."
You don't have a bot and you don't have an agent unless you have a learning loop.
A learning loop is a process where you define an outcome, it generates the results, and at one point there is feedback.
That feedback is fed back into the initial prompt, and then it, it executes with that new knowledge the steps in the process so that it gets better, and it gets better, and it gets better.
So if you have any agents or bots currently in your company and your team is not giving it feedback on its work, then you don't truly have an ability to get better over time.
And that is how the best companies are scaling with AI, by understanding that an agent's not an agent without a learning loop and creating the process.
A simple example of this is having your agent do some kind of analysis, post it into Slack, ask your team for feedback on the analysis, and then tell your agent to monitor when it's doing those posts in Slack, take the feedback and consider that as opportunities to upgrade its brain,
its rules, its steps, its process, its quality.
And then if your team's looking at the output and giving it the feedback and you're feeding that to the agent every week, 52 times a year, it can get better and better and better.
That also informs your data equity. Do you see how these all stack on top of each other?
When we know that AI today is just still allowing most business owners to do the wrong thing or the dumb thing faster, and that we have to develop our skills and understanding this over our team, and our team can never outgrow our ability because we are the leaders or the lids,
and that we need to get better at defining the roles of our agents, and we have to understand the data side of the agents that are, they're going to use that data, and is it good, and do we have great examples of it, and do we know what a win looks like,
and do we know what success looks like, and then build those learning loops into our processes, that is how we make a lot of money.
That's how you win, and it is no longer optional.
2027 will be the separator between the companies that decide to go pro and the amateurs that eventually go out of business.
And when we think about every company now becoming an AI software company, it's not SaaS, it's not software as a service.
It's your service as a software. It's the problem you solve to the market, get really good at that messaging, and then on the back end, use this to deliver at a higher rate, at a higher quality, so that you increase your gross margin because you don't have to hire a bunch of people.
You keep the people doing the people stuff, you use the AI to do the computer stuff, and all of a sudden you make a lot of money.
Real quick, you've made it to the end. Let me tell you what I know about you.
You're not somebody that's looking to do shelf help.
You're not somebody just doom scrolling on social media.
You're here because you want to build, build a real business, and you're sick and tired of not getting the real insights to grow inside your business.
So I want to help you avoid all the pain trying to figure this out the hard way like I did it.
That's why I put together everything I've ever learned into the Scale Workbook.
It's the internal document that I use with every founder I work with that turns a chaotic business into a money-making machine.
And teaching you all of this without burning more hours. It's totally free.
The link is in the description, or you can just scan the QR code on the screen because I'm here to help you build your empire.
TranslateMom transcribes and translates videos from X, TikTok, YouTube and Instagram. Paste a link and you get a page like this one, in your language.