AEO完整工具包（AEO第二部分：现场、非现场与指标）

欢迎收听《早期采用者》，这是一档助你将AI转化为不公平优势的播客。我是杰斯，白天是营销主管，其余清醒时刻则是个AI实验狂人。

Welcome to Early Adopter, the podcast where we help make AI your unfair advantage. I'm Jess, marketing exec by day, AI experimenter by day, night, and, basically every other waking moment in between at this point.

我是凯尔，你们的技术导师。过去十年我一直在构建解决实际问题的数据系统、产品和AI工具。每周我们会解析真正对企业有价值的AI概念、工具或趋势。

And I'm Kyle, your resident tech guru. Spent I the last decade building data systems, products, and AI tools that solve real world problems. Each week, we'll break down AI concepts, tools, or trends that actually matter for your business.

无论你是小企业主、创业者，还是科技爱好者，让我们开始探索吧。就把我们当作你穿越AI这个狂野、奇妙又时而令人窒息的世界的向导。

So whether you're a small business owner, an entrepreneur, or you're just tech curious, let's dive in. Just think of us as your guides through the wild, wonderful, and sometimes completely overwhelming world of AI.

这里是《早期采用者》。

This is Early Adopter.

欢迎回到《早期采用者》，这档助你将AI转化为不公平优势的播客。我是杰斯。

Welcome back to Early Adopter, the podcast where we help make AI your unfair advantage. I'm Jess.

我是凯尔。那么杰斯，先说说你这一周过得如何？

And I'm Kyle. So, Jess, start us off. How was your week?

你可能听出我声音不对劲——我感冒了，这其实是我很久以来第一次感冒。所以提前为...你知道的...所有状况道个歉。

Well, you might be able to hear it in my voice that I have managed to pick up a cold, which is actually the first cold that I've had in ages. So apologies in advance for, you know, everything.

我感觉你和整个英格兰都生病了。比如，我知道在英国我经常交谈的每个人。

I feel like you and all of England has gotten sick. Like, I know every other person I have I speak to on a regular basis in The UK.

是啊。不过，对。这就是我这一周的真实写照，就是工作，还有感冒。这就是我的周。你这一周过得怎么样？

Yeah. But, yeah. So that has been my week, really, just working and, having a cold. That's my week. How's your week been?

嗯，你知道，尽管我一直内向，但这周我却像个社交蝴蝶一样活跃。

Well, you know, despite my constant introversion, I've been quite the social butterfly this week.

哦，瞧瞧你。

Oh, check you out.

我知道。我就是最近特别受欢迎。你知道，我能说什么呢？嗯，我去了，比如，上周的一个万圣节电影之夜

I know. I just I'm very popular these days. You know, what can I say? Well, I've been to, I went to, like, a Halloween movie night last

晚上哦。

night Oh.

有人做了些小点心，比如南瓜和坚果之类的，我当时就想，好吧，现在正式进入秋天了。

Where someone made these, like, little things with, like, pumpkin and nuts and stuff, and I was like, okay. This is now officially fall.

就像，秋天到了。秋天。是的。没错。

Like, it autumn. Autumn. It is. Yes.

我把秋日薄脆饼放进嘴里，美味极了。你知道吗？

I put the autumn cracker in my mouth, and it was delicious. So you know?

你是不是一杯南瓜香料拿铁啊，小妞？

Are you are you a pumpkin spice latte, girly?

我不是，因为我不喜欢糖分，但我确实爱南瓜，比如南瓜派。

I am not because I don't like the sugar, but I do love pumpkin, like pumpkin pie.

哦，南瓜。对。我超爱南瓜派。这很美国，但我确实爱南瓜派。

Oh, pumpkin. Yeah. I do love pumpkin pie. It's so American, but I do love pumpkin pie.

我知道。真正令人难过的是我们一年只能吃一次，但也许好事太多反成负担。如果全年都能吃到，我们可能就不会这么珍惜了。

I know. The fact that the the real sadness is that we only get it once a year, but then maybe it's, like, too much of a good thing. You know, if we had it all year round, maybe we wouldn't really appreciate it.

确实。但你知道吗，十一月之外你也可以享用南瓜风味的东西。这是被允许的。你有自由意志。

True. But, you know, you can have pumpkin spice things outside of November. It is allowed. You have free will.

你知道，我是成年人了，所以我能为所欲为

You know, and I am adult, so I could do whatever I You are

成年人。你当然可以为所欲为。你倒提醒我了，这周我还真做了件和万圣节电影相关的事——这周一刚参加了每月电影知识问答。我知道

an adult. You can do whatever the hell you want. You actually just reminded me that I actually also did a Halloween movie related thing this week. It was my monthly film quiz on Monday this week. I know

电影问答

film quiz.

我知道。你以前可是问答队成员，这周我就在想...

I know. And you used to be on the film quiz team, and this week was I the just think I'm

至少还算是荣誉会员

still like an honorary member of

你当然是。你依然——

my You definitely are. You're still-

我还在WhatsApp群里呢

I'm still on the WhatsApp group.

确实如此。

That is true.

你还在用WhatsApp联系他们。偶尔会这样。

I You're give still on them the WhatsApp time. Every once in a while.

没错。确实如此。但这是因为他们下个月十月不举办活动，因为十一月是他们第二百次问答比赛。所以他们今年提前庆祝了万圣节。

Exactly. You do. But it was because they're not having it next month in October because November is their two hundredth quiz. So they did Halloween early this year.

所以那些恐怖内容很适合，你知道的，恐怖知识问答。我是说，几年前我参加过几次。

So it was horror good content for the horror, you know, the horror trivia. I mean, I had a couple years back.

他们真的全力以赴，这也是他们最喜欢的类型，所以在恐怖问答上他们真的下足了功夫。这是我们举办过最有趣的问答之一，虽然题目真的很难。他们不仅题目出得难，还找了一些最搞笑的恐怖电影片段。他们无所不知。

They they really go it it's their favorite genre as well, so they really go all out on the horror on the horror quiz. And it was one of the funniest quizzes we've had in a A, it was really hard. It was a really, really difficult quiz, but they also, found some of the funniest horror movie clips. They know everything. They know everything.

获胜的团队在60分满分中拿了57分左右。太疯狂了。这些人简直不可思议。令人印象深刻。我知道。

The team that won got out of 60, I think they got 57 or something. That's wild. These people are insane. Impressive. I know.

我们可以整天聊电影问答，确实应该这样

We could talk about Film Quiz all day long, we should

我们或许该言归正传了。

We probably talk probably to the topic at hand.

我知道，我知道。虽然我们可以整天讨论这个。总之，这周我们要完成

I know. I know. Although we could talk about this all day. Anyway, this week we are completing the

这转折真生硬啊，杰西。太生硬了。我

What a segue, Jess. What a segue. I

尽力了。听着，我现在吃了很多感冒药。这周我们将完成关于答案引擎优化的系列第二部分。而且本周我们确实收到了非常及时的新闻，正是关于这个话题。OpenAI这周发布了一个重大公告，基本上验证了我们上周讨论的所有内容。

try. Look, I'm on a lot of cold drugs right now. This week we're completing the second part of our series on Answer Engine Optimization. And we actually had some really timely news this week, on this very topic. OpenAI made a massive announcement this week, basically validating everything that we had talked about last week.

所以事情发展得比我们预想的还要快。

So it's actually happening even faster than I think we expected.

快多了。而且感觉一切都在加速，有点让人应接不暇。

Way faster. And it feels like everything keeps getting faster and slightly overwhelming.

是啊，确实如此。

Yeah, it really is.

但在深入讨论之前，我想我们该快速回顾一下上周内容，为可能错过的人做个简要重述，或者说给大家来个快速复习。我觉得这很有帮助。

But before we dive into that, I think let's do a quick recap for anyone who might've missed last week, or just, you know, let's just do a quick refresher for everyone. Think that's helpful.

虽然我们强烈建议你回听上周的节目，因为如果这周直接听可能会有些地方不理解，不过我们会——

Although we definitely recommend that you go back and listen to last week's episode because some of it might not make sense if you listen this week, but we'll-

是啊，毫不掩饰地打个广告。

Yeah, shameless plug.

上周我们讨论了答案引擎优化（AEO）。它本质上是针对ChatGPT、Claude、Perplexity或谷歌AI模式等人工智能引擎的SEO（搜索引擎优化）。但根本区别在于：传统SEO的目标是排名第一以吸引用户点击进入网站，而AEO则是为了让AI在回答中直接引用你的内容。这是相当不同的。

So last week we covered Answer Engine Optimization or AEO. It's essentially SEO, search engine optimization, for AI engines like ChatGPT or Claude or Perplexity or Google's AI mode. But the fundamental difference is that traditional SEO was all about ranking number one so that people would click through to your website. AEO is about getting cited in the AI's answer itself. So it is quite different.

这之所以重要，是因为相关数据开始变得非常显著且值得关注。正如上周所说，现在60%的搜索行为最终都不会点击进入网站。实际上HubSpot的最新研究也验证了这点——人们直接在AI界面获取答案，完全不再访问网站。

And the reason this matters is the numbers are starting to become really stark and really important to keep in mind. So as we talked about last week, 60% of searches now end without a click to your website. And actually, that now has been validated by HubSpot's recent research as well. People are getting their answers directly in the AI interface and never visiting websites at all.

而当用户确实通过AI推荐点击进入时，转化率会高出非常多。我们可能会多次提及Webflow这个案例：他们的LLM流量转化率比传统谷歌搜索流量高出六倍。六倍不是边际改善，而是真正的变革性差异。

And when people do click through from an AI recommendation, the conversion rates are much, much, much higher. So an example that we'll probably talk about a fair bit is Webflow. Webflow saw a six x conversion rate difference between LLM traffic and traditional Google search traffic. Six times is not just a marginal improvement. Like, that is actually transformational.

没错。上周我们也分析过，这很合理——因为当用户真正点击进入你网站时，他们已经完成了完整的个人探索旅程，相当于和AI进行了咨询。他们已明确痛点、寻求过建议、做过调研。如果他们按照我们说的提示词操作，其实已经历多次迭代才最终找到你的公司或网站。

Yeah. And we kind of walked through last week. It makes a lot of sense because by the time someone actually clicks through to your website, they've gone through a whole kind of personal discovery journey or like a consultation with the AI. They've asked for they've identified their pain point, they've asked for potential recommendations, they've researched it. If they're doing all of the prompting that we talk about, they've kind of gone through quite a bit of iteration to then get to your company or your site.

AI非常了解他们的具体需求。我们谈论的漏斗，比传统营销漏斗深入得多——客户已过了发现阶段，甚至可能已经预先筛选过，所以这是非常优质的潜在客户。

And AI understands a lot of their specific needs. So what we talk about, like, with funnels, there's so much further down the funnel than the typical marketing funnel where they're just in the discovery phase. They're prequalified or maybe even more than that. So it's a really good lead.

作为回顾，这就是我们讨论的背景。现在让我们进入战术层面，因为上周我们重点讨论了其重要性，而本周我们要真正探讨如何实际操作。我将基于我们的研究，将其分解为一个框架。特别是莱尼播客中对Graphite公司Ethan Smith的采访。

So that is the context as a recap of what we're talking about. So let's get tactical because last week we really dedicated to just talking about why it was important. And this week we really want to get into how do you actually do this. So I'm going to break this down into a framework based on some research that we've done. In particular, there's an interview with a guy called Ethan Smith from Graphite, which is on Lenny's pod.

如果你还没听过，这可能是目前关于AEO最全面的战术指南。他免费分享了大量宝贵信息，绝对值得一听。我们会确保在节目备注中附上链接，方便你回听完整内容。

If you haven't listened to it, it's probably the most comprehensive tactical guide on AEO that exists right now. He gives away an incredible amount of information for free. So that is definitely worth listening. We'll make sure that that is in the show notes so you can go back and listen to it in full.

主要有三大支柱：首先是站内优化，其次是站外引用建设，最后是效果衡量。但在深入之前，杰西，我们需要先明确几点——这绝不是要钻系统空子、制造垃圾信息，或者像我常说的那种AI销售机器人，机械地海量触达用户。

There are essentially three main pillars. The first one, on-site optimization, then you have off-site citation building, and then lastly, measurement. But I think before we dive into that, Jess, we should probably address some important things. This isn't necessarily about gaming the system or creating spam or as I always like to talk about being another, like, AI SDR bot that's gonna reach out to a million people for you.

没错，感谢你提到这点。稍后我们会详细讨论，但必须重申整个方法必须植根于真实性和真诚助人的初衷，提供真正有价值的内容——这正是当前SEO走偏的地方，不过这个话题我们稍后再展开。

No. And thank you for saying that. And we'll talk more about this later, but it's really important to reiterate that the entire approach really has to be rooted in authenticity and genuinely helping people and providing people with content that is genuinely useful, because that is kind of where SEO has gone a little bit wrong. But we'll we'll get more into that later.

我认为这正是它对整个生态系统和营销领域大有裨益的原因。如你所说，传统SEO可以靠关键词堆砌或重金购买核心词排名，而AEO奖励的则是真正有用、清晰、符合用户需求的内容。

And I think this is one of the reasons it could be really good for the overall ecosystem and for marketing as a whole. As you pointed out, unlike traditional SEO, where you could keyword stuff your way to the top or spend a lot of money to capture those really important keywords, AEO on the other hand is going to reward genuinely helpful, clear content, what you're actually looking for.

秉持这个理念，现在我们来拆解具体操作方法。我们将逐步讲解这三大支柱，首先从站内优化开始。

So with that philosophy in mind, let's break down how you actually do this. And we're going to walk through these three main pillars. So let's start with on-site optimization.

是的。我们从这里开始是因为，你知道，网站是基础。它仍然是每个人直接与你接触时的消费旅程起点。但就像我们上周讨论的，SEO并非突然就不重要了。我们实际上是站在SEO的肩膀上发展AEO。

Yeah. And we start there because, you know, your website is the foundation. It's still the start of, you know, everyone's consumer journey when they're dealing with you directly. But and as we kind of talked about last week, it's not like SEO suddenly doesn't matter. We're we're sitting on the shoulders of SEO with AEO.

没错。所以SEO对AEO仍然重要，因为如果你的页面在传统搜索中完全排不上名次，它们可能就无法进入ChatGPT或其他AI搜索引擎调用的检索池。如果把SEO比作入场券，它算是让你进入房间，而AEO则是让AI真正点你名的关键。

Yep, that's right. So SEO still matters for AEO because if your pages don't rank at all in the traditional search, they might not get into the retrieval pool that ChatGPT or any of the other AI search engines pull from. So if you think of SEO as like an entry ticket, it kind of gets you into the room, but then AEO is what makes AI actually call your name.

是的。我上周也很喜欢这个比喻。

Yeah. I loved that analogy last week too.

嗯，我觉得这个解释确实非常清晰明了。

Well, I think it just it it's a really clear explanation.

没错。完全同意。但杰西，关于AEO与传统SEO在站内优化方面的细微差别，比如反向链接这些，有哪些需要注意的呢？

Yeah. Absolutely. But what are what are some of the nuances, Jess, around on-site optimization for AEO versus traditional SEO with backlinking and all of that?

最大的区别在于长尾词。我的意思是，ChatGPT查询的平均词数约为25个，而谷歌搜索只有6个词左右。人们会提出更长、更像对话的问题。这也是我们讨论对话式搜索的原因之一——它更像是自然的来回对话，而非传统搜索模式。比如用户不会搜索'会计工具'，而是问'适合小型企业、能处理多币种且无需金融学位就能操作的好用会计软件有哪些'。

So the biggest difference is the long tail. And by that, I mean the average number of words in a ChatGPT query is around 25 words versus Google, which is around six words. So people are asking much longer, more conversational questions. And this is one of the reasons why we talk about conversational search, because it's that kind of back and forth, like a natural conversation versus what we traditionally have used for search. So instead of a search query like accounting tool, they're asking, what's good accounting software for a small business that needs to handle multiple currencies and it doesn't require a finance degree to use.

确实。传统关键词研究工具捕捉不到这类查询，因为其中许多问题从未在谷歌上被搜索过。所以对话式搜索的长尾词规模是极其庞大的。

Yeah. And traditional keyword research tools don't capture those queries because many of them have never been searched on Google before. So the long tail is enormous with conversational search.

那么我们现在把这个放入一个实际的框架中，因为我们喜欢框架。第一步，我们称之为问题挖掘。你需要识别人们实际上对你的产品或服务提出了哪些问题。有多个来源可以查阅，但这本质上能让你在创建内容时了解可以解答哪些问题。

So let's kind of put this into an actual framework now, because we love a framework. So step one, let's call it question mining. So you need to identify what questions people are actually asking about your product or your service. And there are several sources that you can look at for that. But essentially what this allows you to do is to start to get an idea of what questions you can address when you are creating content.

因为否则，如果你不知道问题会是什么，你怎么知道应该回应什么？

Because otherwise, how do you know what you should be responding to if you don't know what the questions are going be?

是的。其中一些来源包括存在于你网站之外世界各地的所有问题。这些可以是销售电话、客户支持工单、Reddit论坛，你也可以去G2或Gartner等地方，那里人们会给出评论和对业务的整体看法。

Yeah. So some of those sources are all the questions that exist somewhere out in the world outside of your website. So these can be sales calls, customer support tickets, this can be Reddit forums, you could go to G2 or Gartner or some of these places where, you know, people are giving reviews and kind of overall views of your business.

没错，实际上你的销售电话对这类事情可能是金矿。这就是为什么我们讨论的那80%非结构化数据如此重要。

Yeah, and actually, your sales calls are probably gold for this sort of thing. This is why it's really important to have this unstructured data that we talk about in our 80%

你的数据中有80%是非结构化的。

of your data is unstructured.

完全正确。我们在数据相关节目中讨论过，这正是最适合提取的数据类型，因为它会让你了解潜在客户反复提出的问题、他们提出的反对意见、询问的功能以及你经常遇到的问题。

Exactly. And we talked about it in our data episodes, but this is exactly the sort of data that is perfect to to pull from because it's gonna give you an idea of what questions prospects ask prospects ask repeatedly, what objections do they raise, what features do they ask about, what issues are you constantly coming across?

是的，客户支持工单也是个金矿。你知道人们困惑什么？他们在哪些地方遇到困难？他们无法解决什么？他们需要什么帮助？

Yeah, customer support tickets are a gold mine too. You know, what are people confused about? What are they getting tripped up on? What are they unable to solve? What do they need help with?

如果你有用于收集这些信息的客户支持工具或网站上的聊天机器人，就去查看那些日志记录。这些通常是非常具体的长尾问题，同样非常适合以某种ChatGPT形式重现。

So if you have a customer support tool that you use to capture all of this or a chatbot on your site, go through those logs and look at that. These are often very specific long tail questions that are again, are going to be perfect for recreating in some form of ChatGPT.

当然还有Reddit、行业论坛等人们会自然讨论的地方——抱歉，应该说任何人们会自然讨论你产品所解决问题的平台。这些都是好去处，特别是Reddit。我们稍后会详细讨论Reddit。

And then of course, you've got your Reddits, your industry forums, anywhere that people will naturally discuss. Sorry, can't speak anywhere where people will naturally discuss the problems that your product solves. That's going to be a good place to go because especially Reddit. And we'll talk more about Reddit later.

另一个好技巧是将现有SEO关键词转化为问题。快速方法是把它们交给ChatGPT，让它基于你已购买或已被证明对业务有价值的关键词，生成20个人们可能会问的问题变体。它在这方面确实很在行，因为它理解人们如何自然地组织对话式查询。

I think another good trick is take your existing SEO keywords and transform them into questions. So a quick hack, give them to ChatGPT, ask it to generate 20 variations of questions people might ask based on the keywords that you're already buying or the keywords that have proven to be really valuable for your business. It's actually really good at this because it understands how people naturally phrase conversational queries.

完成问题挖掘后，我们进入第二步：创建全面的可引用内容。这正是该方法与传统内容营销显著不同之处。每个页面应该针对数百个问题，而不仅是一个关键词。回答的问题越多越好。

So once you've done your question mining, then we move on to step two, which is where you create comprehensive quotable content. Content. And this is where the approach really differs significantly from traditional content marketing. Each page should target hundreds of questions, not just one keyword. So the more questions you answer, the better.

我知道这看起来工作量很大，但它确实非常重要。

And I know this seems like a lot, but it is actually very important.

这是个绝佳的用户旅程演练。想象你正在和某人讨论产品，当他们提出一个问题时，你会回答，然后他们自然会追问三四个后续问题。你的内容应该预判并回答所有这些潜在问题。

It's a great, like, customer journey exercise to go through. You know, think of it like you're having conversation with someone about your product. When they ask you one question, you're going to answer that. And then they're naturally going to ask you three or more follow-up questions. So your content should anticipate and answer all of those potential questions.

以项目管理工具为例，假设我们是monday.com。你的页面不应该只说'我们是项目管理工具'，这实际上没有回答问题。它应该回答：你们有哪些功能？支持哪些集成？

So if you look at like a product management tool example, so something like, let's pretend we're monday.com, right? Your page shouldn't just say we're a project management tool. That actually doesn't really answer the questions. It should answer what features do you have? What integrations do you support?

你们的定价如何？与竞品相比有什么优势？这里有一长串问题需要解答。不过话说回来，这些都是企业在发展过程中必然会遇到的疑问。所以这些问题是真实存在的。

What's your pricing? How does it compare to alternatives? There's quite a long list of questions that it should be answering. But again, these are all questions that you will have encountered at some point within your company's lifespan. So it exists.

关键就在于找出这些问题。

It's just about finding those questions.

真正关键的是，如果你真的着手做这个练习，务必确保用清晰的事实陈述直接回答问题。不要用大量营销术语和让人摸不着头脑的话——我在浏览网站时经常看到这种内容，完全不知所云。应该这样说：'支持与2000多款应用集成'，然后列出前10或20个，比如Slack、Gmail、Salesforce等。另一个好例子是'可在低至华氏15度的环境中正常工作'。虽然不确定这个例子适用于什么产品，但对AI来说引用这种具体描述会非常容易。

And the really critical part, if you do go through and do this exercise, you want to make sure you have clear factual statements that directly answer questions. So not a lot of marketing buzzwords and things that don't really make sense, which I show up on websites and I go, what are they talking Right? But things like, you know, supports integrations with over 2,000 apps including, and then list out, like, the top 10 or 20, Slack, Gmail, Salesforce, whatever it may be. Another good example is works in temperatures down to 15 degrees Fahrenheit. If you're I'm not sure what example that would be good for, but let's try it would be really easy and really specific for an AI to quote that.

凯尔，我觉得如果你卖的是冬季外套这类产品，这个例子就非常贴切。比如羽绒服。这就对了。这就是个很棒的产品描述范例。

Well, I think you'll find, Kyle, that that would be great if you were selling a winter coat, for example. That is a bag. A There you go. So that's a great product description example. Exactly.

所以你看，如果只说些模糊的话，比如'我们能与所有你喜爱的工具集成'，这对AI毫无用处，因为AI需要回答具体问题。而且我们此刻使用的自然语言对话模式对AI至关重要——这正是AI的交互方式。要用和人交谈的语气写作，而不是像为了搜索引擎算法堆砌关键词。现在很多网站都深受其害。AI需要能像对话一样解析理解内容，而不是仅依赖关键词。

So, you know, just saying something really vague, like we integrate with all your favorite tools. That's really useless for an AI because the AI needs to answer a specific question. And also natural language, which is the way that we are talking back and forth right now is critical for AI, because this is how AI interacts. You know, write like you're talking to a person, not like you're just stuffing a website with keywords for the search algorithm, which is something that a lot of websites actually really suffer from right now. You know, the AI needs to be able to parse and understand your content conversationally rather than just keyword based.

这就是我对AEO世界充满希望的原因。说真的，你们有多少次点开网站看到的全是营销话术？本质上都是在堆砌关键词搞SEO优化。但当我用对话方式阅读时，根本看不懂你们在说什么。

This is why I have so much hope for this AEO world because, I mean, it's true. I how many times do you go to a website and you read the descriptions and it's marketing buzzword, but what it really is is keywords. They're keyword farming because they're trying to basically stuff the litter box with high SEO ranking. But when I read it conversationally, I have no clue what you're doing. Right?

最后我不得不去翻解决方案、产品和案例研究页面，试图搞明白：这到底是干什么的？而在新的AEO世界里，人们终于会说人话：'我们做这个，我们集成那些'，用真实的语言而不是晦涩术语——就像给五岁小孩解释那样。我真心希望世界能转向这种模式。

And I have to, like, go to solutions and products and case studies, and I have to try to figure out, okay, what the heck is this thing? Yeah. Whereas I think in this new AEO world, you're finally gonna have people saying, like, we do this thing. We integrate with those things, and you're actually gonna have, like, real words, not this buzzwordy, like, explain like I'm five. I would love the world just to switch towards this.

我认为AEO可能会推动这种转变。别忘了，结构化数据依然重要。

I think AEO is gonna maybe force that shift. And don't forget, structured data still matters.

是的。清晰的标题。非结构化数据很重要，嗯哼。结构化数据也是。

Yeah. Clear headings. Unstructured data is important, but Mhmm. So is your structured data.

没错。所以要有清晰的标题、适当使用项目符号，如果可能的话加上模式标记，还有数字、统计数据、明确的投资回报率数据，这些都非常有帮助，因为很可能有人会问：这个的投资回报率是多少？

Yes. So clear headings, bullet points where appropriate, and the schema markups if you can implement that, but also numbers, statistics, clear ROI figures, super helpful because you're probably going to have people asking, what is the ROI of this?

AI特别喜欢统计数据，这就像是它总在幻想的东西之一。所以能有一个真实可靠的统计数据真的很有用。但我们能不能简单解释一下？我要按下术语警报器了。我们能不能聊聊什么是模式标记？

AI loves a good statistic, and it's like one of the things that it's always hallucinating. So to actually have a legit a legit stat is really useful. But can we just quickly demystify? I'm going to buzz my jargon buzzer. Can we just talk about what a schema markup is?

因为我觉得有些人听到这个词但并不真正明白它的意思。那么什么是模式标记呢？

Because I think some people hear that don't actually know what it means. So what is a schema markup?

好的，说得对。感谢你的警报器。其实并不复杂。模式标记就是给你的内容加上标签，告诉机器或爬虫这是价格、这是产品名称、这是库存状态。你基本上是在给它一些小标记让它理解上下文。

Yeah, fair enough. I appreciate the buzzer. And it's not super technical. A schema markup is just adding labels to your content that tells the machine or the crawler, this is a price or this is a product name, this is availability. So you're kind of just giving it little markers so that it knows the context.

那些在Google上显示的食谱卡片，带有烹饪时间和配料信息的，就是模式标记的实际应用。

Those recipe cards that show up in Google with like cooking time and ingredients, that's schema markup in action.

好的。那么，这一切意味着你不是让AI去猜测信息是什么，而是明确地进行标注。是这样吗？

Okay. So, all of this means that instead of having your AI having to guess what information is what, you're explicitly labeling it. Is that right?

没错。如果你使用Shopify或WordPress作为网站构建工具，很多功能已经内置其中。它们会为你处理结构化标记。如果是定制网站，开发者可以添加这个功能。你不需要成为专家，只需知道它的存在，并确保开发者将其纳入他们创建的内容中。

Exactly. So if you're on Shopify or WordPress for your website builder, that's a lot of that is built in. They will do the schema markup for you. If you have a custom site, your developer can add it. You don't need to become an expert, you know, just know it exists and make sure that your developer is building that into things that they create.

所以直接问他们：‘我们设置了这个吗？’老实说，我认为最重要的是针对人们实际提出的问题，写出清晰、全面的回答。结构化标记有助于告诉后台机器这是什么，但你知道，如果内容本身足够清晰，这才是最关键的。

So just ask them, do we have this set up? I think honestly, the most important thing is just writing really clear, comprehensive answers to real questions people have. The schema markup is helpful to to tell the machine in the background what it is, but, you know, it's kind of a a hack if clear content is gonna be the most essential.

记得我们之前在做关于提示工程的节目时说过，提示工程本质上就是良好沟通的华丽说法。我认为这再次归结到这一点，对吧？这是完全相同的概念。而且这似乎是AI的一个永恒主题——一切都关乎清晰的沟通、良好的简报，给予它所需的背景和上下文，因为这就是AI的工作原理。所以，是的，我认为这非常关键。

So you know how we we said a while back when we were doing our episodes about prompt engineering, that really prompt engineering is just fancy wording for good communication. And I think, again, that's what it boils down to, right? This is the exact same concept. And I think this is kind of a running theme with AI, is that it's all about clear communication, good briefing, just giving it the background and the context that it needs, because that is how the AI works. So, yeah, I think that's really key.

结构化标记是锦上添花，但你说的沟通清晰才是必须的。

And the schema is a nice to have, but what you're saying, clarity in your communication is the must have.

正是如此。

Exactly.

好的，那么你下一步的关键是优化帮助中心。这一点让我有些意外，但细想完全合理。帮助中心或知识库对AEO至关重要，因为人们会向AI引擎提出非常具体的问题，比如产品是否能满足需求。例如询问具体任务或功能：‘这款CRM能把数据导出到BigQuery吗？’

Okay, so your next critical step is to optimize your help center. This is something that kind of surprised me, but I think makes total sense. Your help center or your knowledge base is critical for AEO because people are asking these AI engines really specific questions about whether products can do what it needs to do. So asking very specific tasks or functions. Can this CRM export data to BigQuery?

对吧？这是个非常具体的问题。你通常只会在知识中心里回答这类专门为某人解答的问题。或者说，这款会计软件能处理多币种交易吗？这类内容传统上一般不会放在官网页面上。

Right? That's a very specific question. You're really only gonna answer that in your knowledge hub where you've tried to answer that for someone specifically. Or, you know, does this accounting software handle multi currency transactions? You probably haven't put that on your website traditionally.

这些信息可能藏在技术文档的某个角落。但如果无处可寻，那它就不会出现在ChatGPT的搜索结果中。

That's probably sitting somewhere in technical documentation. But if it's not sitting somewhere, then you're not going to show up in the chat GPT result.

我认为这类细粒度问题通常在帮助中心或FAQ里能找到答案。这类信息往往存在于常见问题解答中，而不是营销描述页面上。在这个特定场景下，你可能成为解答那个非常冷门问题的唯一来源，这意味着你可以完全掌控这个搜索查询。

I think these granular questions are often answered in help centers or in FAQs. And that sort of information often sits in those sorts of FAQs, not on the marketing description page. And that might be the only, you, you in this particular instance might be the only source answering that really specific obscure question, which means that you could really own that query entirely.

以上就是第一支柱，它非常聚焦于你的网站。接下来我们要讨论的第二支柱是站外引用建设，这里正是AEO开始与传统SEO分道扬镳的地方。我觉得这很令人兴奋——传统SEO模式下，初创企业因缺乏域名权威而难以竞争，也可能没有足够的预算来购买关键词等资源。

So that was the first pillar, and it's very focused around your website, right? This second pillar that we're going get into is off-site citation building, and this is really where AEO starts to diverge from traditional SEO. And I I think it's kind of exciting. So with traditional SEO, early stage companies have had trouble competing because they don't have domain authority. They may not have the same budgets for, you know, keywords, what have you.

但通过AEO，你可以像我们讨论的那样出现在Reddit话题中，或被其他高权威网站提及，从而立即出现在AI答案里。这确实是实现曝光民主化的绝佳机会。

But with AEO, you can get mentioned in a Reddit thread like we talk about, or you can be talked about, you know, elsewhere on high domain authority websites and start appearing in AI Answers immediately. So it's a really great opportunity to democratize mentions.

你要不要快速解释下什么是域名权威？因为我觉得很多人可能不清楚。我要临时考考你这个问题，抱歉啦。

And do you want to quickly explain what domain authority is? Because I think a lot of people might not know. I'm going put you on the spot here and ask that question. Sorry.

好的。如果要我总结的话，域名权威本质上是个SEO指标（历史上如此），它用于预测或数学量化网站在搜索引擎结果中的排名潜力。虽然它不直接作为谷歌排名因素，但对理解网站整体实力很有帮助。通过获得大量权威网站的反向链接，就能建立高域名权威。

Yeah. Let me if I was going to summarize domain authority, it's really a SEO metric, kind of historically, and it's the ability to predict or mathematically quantify a website's potential to rank in search engine results. But it's really useful for, you know, understanding a site's overall strength, And it's not necessarily a Google ranking factor. Like, it doesn't directly impact your rankings. But a high domain authority can be built by earning, you know, a lot of link backs from other authoritative websites.

这是谷歌确认你是一个真实网站的方式。

It's it's a way of Google understanding that you're a real website.

这可以说是另一种更简单有效的方法。但本质上它只是让谷歌知道你不是垃圾网站或虚假网站。之所以有权威性是有原因的。这就是它的本质。

That's kind of the other, the easiest better. But it basically just allows Google to know that you're not a spam site or that you're not a fake website. There's a reason why it's authority. So that is what it is essentially.

它并不直接关联到显示为第一个搜索结果，但更高的域名权威性意味着你的排名会更靠前。

It's not directly linked to showing up as the first result, but a higher domain authority will mean you rank higher.

这一点很重要，因为实际上这是小型企业或初创公司可以通过AEO真正获胜的地方。SEO和域名权威性通常需要投入大量资金或专业知识，尤其是现在——要知道二十年前SEO刚开始发展时可能不是这样。但现在要在谷歌第一页排名非常困难，因为这些大公司基本垄断了那个空间。而AEO仍处于早期阶段公司可以获胜的阶段，你只需要在Reddit话题中被提及就能快速获胜，特别是在那些尚未有太多讨论的专业领域。这比经历昂贵且复杂的域名权威SEO过程要容易得多。

So this is an important point because this is actually where smaller businesses or startups can really win with AEO because SEO and things like domain authority often need a lot of money or a lot of expertise thrown at them, especially now, you know, twenty years in, maybe not twenty years ago when SEO was kind of first starting to get off the ground. But now it's very difficult to rank on the first page of Google. That is because these big companies kind of own that space. But with AEO, we're still in the phase where earlier stage companies can win and you can win pretty quickly just by getting mentioned in a Reddit thread or especially if it's an area of expertise where there might not be a lot of conversation about it yet. So it's a much easier, win than having to go through the whole domain authority SEO process, which is expensive and complicated.

是的，

Yeah,

完全同意。关键区别在于，在谷歌或传统方式中，如果你的网址、如果你的网站排在第一位，你就赢了。但在LLM中，LLM是基于其信息池汇总多个引用来源。所以你需要尽可能多地被提及。关键在于这些来源中的引用频率，这一点非常重要。

completely agree. The key difference here is that in Google or traditionally, if your URL, if your site showed up first, I mean, you want you win. But in LLMs, the LLM is summarizing many citations based on its pool of information. So you need to get mentioned as many times as possible. It's about frequency of citation across those sources, which is super important.

我认为还有一点需要注意，你越早这样做，出现得越多，人们反馈说这是个相关结果，这种‘你是他们的正确结果’的强化作用以及他们点击的行为，意味着你会越来越频繁地出现。因此你也可以通过这种强化机制在AEO中建立相当的权威性。这里有几个例子。我们特别来看看Reddit，因为我们经常讨论它。

And I think also one thing to kind of note here, the sooner you do this and the more you show up, and people kind of giving feedback saying this is a relevant result, that kind of reinforcement around you were the right result for them and they click on that means you're going to now show up more and more. So you can build quite a bit of authority within the AEO from that reinforcement mechanism as well. A couple examples here. Let's let's take a look at Reddit in particular because we talk about that a lot.

是的，我们确实如此。Reddit对AEO（作者体验优化）至关重要。对于大多数大型语言模型来说，依赖用户意见的第三方内容比品牌生成的内容更有分量。ChatGPT、谷歌的AI模式等都非常依赖Reddit和LinkedIn这类平台。

Yeah. We do. Reddit is huge for AEO. Third party content, which relies on user opinions, carries more weight than branded than brand generated content for most LLMs. ChatGPT, Google's AI mode, all of those rely really heavily on sites like Reddit and actually LinkedIn as well.

LinkedIn也是个重要平台，是另一个可以建立行业信誉的地方。

LinkedIn's a big one too, so that's another place where you can go and start to earn some street cred.

但我觉得多数人犯错的地方在于试图走捷径，比如采用增长黑客策略——创建数百个虚假Reddit账号、到处自动发评论，试图钻系统空子。

But I think where most people get it wrong is they try to hack it or, you know, they try that growth hacker approach where you create hundreds of fake Reddit accounts, you auto post comments everywhere, and you try to game the system.

这在Reddit上尤其行不通。任何发垃圾信息的人都会立刻失败：账号会被封禁，评论会被删除，基本上会被赶出Reddit。

That is not going to work on, especially on a place like Reddit. You know, anyone trying to spam is going to immediately be unsuccessful. Your account's going to get banned. Those comments are going to get deleted. You're basically gonna get run off of Reddit.

Reddit非常...是的，他们非常清楚这类把戏，所以根本没用。

Reddit is a very, Yeah. They're very aware of those sorts of tactics, so it's not gonna work.

这绝对会损害声誉。我见过太多帖子，有人发完内容后就被网友集体嘲讽。你可以查看账号创建时间和历史发帖记录...

Yeah. It definitely adds to reputational damage. I've seen it in so many different threads where someone will post something and then you just have this long chain of people roasting the user. They're like and and they'll you can look, right? So you can see when that user was created and there are other posts.

一看就知道是广告机器人。这样你反而会彻底失去想触达的受众。真正有效的策略很简单：找到目标引用话题，坦诚说明身份和工作单位，提供有用信息。这本质上就是保持内容真实性的核心——你传递信息的可信度。

So you can look at that and go, oh, obviously that's an ad bot. And then you completely lose all those people that you were trying to reach. So the strategy that works is really simple. Find a thread that's part of a citation you want to show up in, say who you are, say where you work, be honest and up front, and give useful information. And again, that's all part of that authentic component, like that authenticity of the information you're trying

分享下。在Lenny对Ethan Smith的那期播客访谈中——再次强烈建议回听——他们谈到了客户之一Webflow。Webflow有几位员工会直接去评论区留言，他们非常坦率诚实，会明确说：'这是我的名字，我在Webflow工作'。

to share. In that Lenny's pod interview with Ethan Smith, which again, absolutely go back and listen to, they talk about Webflow, which is one of their clients. And they have a couple people who work at Webflow who go to the comments and they are very upfront and very honest about it. They say, This is my name. I work at Webflow.

这里有个实用信息点。但由于这仍属于真实对话，且他们坦诚公开身份和所属公司，这些信息依然有价值——毕竟人们确实需要了解网站建设相关知识。只要不是暗箱操作，这种真诚分享就会受到欢迎。

Here's a useful piece of information. But because it's still part of that authentic conversation and because they're being honest about who they are and where they work, that's still useful because people still need to know stuff about website building. But as long as someone's not trying to be sneaky about it, then it's going to be welcomed because they are being authentic.

完全同意。我经常看到这种有效方式：'嘿，这是你问题的解决方案，这是我尝试过的方法。顺便说下，我是某某公司员工，我们是这样解决该难题的。'

Absolutely. I see that work all the time where someone says, Hey, here's an answer to your question. Here's something I've tried. Oh, and by the way, this is who I am. This is the company I work for, and this is how we're solving that challenge.

这种内容我的点击率几乎100%。我总会关注这类回复，因为...

That I I almost my click rate on that is probably a 100%. I'm always looking at that because they

他们清楚自己在说什么，确实具备专业认知。

show know what they talk about. They're they know what they're talking about.

他们建立了可信度，提供了实用信息，还告诉我：'我正在这个领域解决该问题'。我就会想：'好，让我去了解下'。

They've shown credibility. They've given a useful piece of information, and they're telling me, hey. I'm trying to solve that problem in the space. I'm like, okay. Let me go look at that.

对吧？我想看看他们具体怎么做。你会想支持这种程度的真诚。事实上根本不需要上万条评论，也不需要大规模刷遍Reddit。

Right? Let me see what they're doing. You wanna support that level of authenticity. And the the truth is you don't need 10,000 comments. You don't need farm all of Reddit at scale.

要知道，你们确实在关注互动质量。所以Reddit的策略就是真正成为Reddit的用户。注册账号，介绍自己，说明工作单位，提供有用的信息和技巧，这样能帮助建立真实性和可信度。

You know, you really are focusing on quality of engagement. So the Reddit strategy is just be an actual user of Reddit. Make an account, say who you are, say where you work, and give useful information and tips, and that will help you build authenticity and credibility.

说实话，社区本就该这样运作。如果你真的在一家解决问题的公司工作，分享这些信息并提供帮助对所有人都有益。这样实际上能带来好处。

And honestly, this is how communities should work anyway. Like, if you genuinely work at a company that's solving a problem that someone has, sharing that information and being helpful is going to be useful for everyone. So that there's actually a benefit to that.

没错。好的。下一个我们要讲的是YouTube。YouTube视频对AEO效果很好。我认为这对B2B尤其有效，因为那个领域的竞争会少很多。

Yep. Alright. The next one we have is YouTube. So YouTube videos work really well for AEO. And I think it's particularly powerful for B2B because there's going to be a lot less competition in that space.

是啊。我是说，关于AI驱动的支付处理API的视频并不多，虽然这个话题很刺激。但从AEO角度来看这是个非常强的关键词，因为如果你为这些非常具体的问题或主题制作视频，机会很大，因为没多少人会讨论这些。所以这更像是个待开发的机会。

Yeah. I mean, there's not that many videos about AI powered payment processing APIs, as thrilling as that is. But it's a really, really strong term from an AEO perspective because if you're making a video for these really specific questions or topics, there's a really big opportunity in that because not many people are going to be talking about it. So it's more of an opportunity to be decided.

对。由于YouTube的规模，你可以找到这些细分受众，为他们解答具体问题，再次获得高参与度。而且这不需要花哨的制作。只需屏幕录制配上语音讲解如何解决问题。我看过很多关于如何构建AI代理的视频，说实话，它们并不炫酷。

Yeah. Because of the scale of YouTube, you can find these niche audiences where you can answer specific questions for those groups, and again, have really high engagement. And you don't need fancy production for this. Just screen recordings with a voice over explaining how to solve problems. I mean, the amount of videos I watch around how to build AI agents from people, and look, they're not sexy.

制作质量不高。

Not high quality production.

确实不怎么样。

No, they are not.

但它们确实信息丰富。最终，AI会从文字记录中提取内容。它们不会在意你的灯光效果或制作水准，这些根本不重要。AI关注的是信息的可信度和实用性。

But they're informative. And, ultimately, that, the AI is gonna pull from the transcript. They're not gonna care about they're they're not looking at your lighting or your production value. They don't care about that. They care about how credible and how useful that information is.

完全同意。比如关于如何将X与Y整合的视频，或是解释A与B区别的内容，以及具体用例的逐步演示。这些都是AI引擎非常、非常、非常喜欢引用的素材。没错。

Totally. So like a video about how to integrate X with Y, or what's the difference between A and B, like walkthroughs of specific use cases. Those are all really, really, really quotable for AI engines. Yeah.

YouTube与Reddit的不同之处在于，它没有社区限制问题。你可以制作视频并随心所欲地使用它，可能有人看也可能没有，但你的作品会成为引用池的一部分。

And the thing about YouTube where it's a little different with Reddit, where you don't have that community issue, you know, you can make a video and do whatever you want with it. You know, maybe people view it, maybe they don't, but you can make it and it becomes part of the citation pool.

好的。接下来特别有用的资源是与出版商的合作关系及原创研究。这意味着要在行业刊物中被提及、撰写客座文章、参与播客节目。所有这些都能扩充你的引用池，因为能增加你的曝光度和发声机会。

Okay. So moving on to another source that is particularly useful is publisher relationships and original research. So this means getting mentioned in industry publications, doing guest posts, participating in podcasts. All of this is going to build your citation pool because it's just getting you out there and talking more.

正如我们讨论过的，AI代理和爬虫注重权威性与可信度。这包括官方网站。我们详细讨论过《纽约时报》的各种合作案例，比如OpenAI这类公司与高权威出版机构的合作，还有维基百科。被众多可信出版物报道和转载会带来巨大影响力，而且这种趋势只会越来越强。

As we've talked about, AI agents and crawlers focus on authority and credibility. So that includes official websites. And obviously we've talked a lot about the different New York Times issues and the partnerships that, you know, companies like OpenAI have with some, you know, high authority publishers, and then also we've talked about Wikipedia. So getting covered and being published in a lot of trusted publications carries huge amount of weight and probably will continue to be more and more.

如果你能产出原创研究或原始数据，那同样价值巨大。无论是白皮书、行业报告、调查还是数据分析，只要发布包含原始数据的内容，AI就特别喜欢引用具体数字。其他文章会开始引用你，这将显著放大你的引用存在感。这一点其实与SEO非常相似。

And if you create original research or original data, that is also huge. So that's a white paper, that's industry reports, surveys, data analysis. If you publish something with original data, there's that AI loves to cite a number. Other articles will start to cite you and then that just kind of amplifies your citation presence significantly. So that is, and that is not too dissimilar from SEO.

但话说回来，关键还是在于你的写作方式——我认为这取决于你如何表述内容。

But again, it's like how you're writing it, I think in the way that you're writing it.

是的，一个很好的例子就是，根据xyz公司2025年的报告，73%的营销人员表示空白。这会在每次聊天中出现，对吧？这具有很高的权威性，非常口语化，而且极具引用价值。

Yeah, think a really good example of this is, you know, according to x y z company twenty twenty five report, 73% of marketers say blank. That is gonna show up in every chat. Right? That is, like, very high authority, very conversational. It's extremely quotable.

人工智能引擎特别喜欢引用这类具体统计数据和数据点。

AI engines love citing specific statistics and data points like that.

没错。我们还应该指出，如果你有钱，可以付费让像《福布斯》这样的网站或出版物提到你，然后突然间你就成了《福布斯》评选的20大最佳信用卡之一。虽然价格昂贵，但也是个选择。我还得说，这就是人们如何登上30位30岁以下精英榜的。听起来很厉害，但那是花钱买的。

Yeah. We should also flag that if you have money, you can pay a site like or a publication like Forbes to mention you, and then, you know, suddenly you're part of the top 20 best credit cards according to Forbes. It's expensive, but it's also an option. I should also say that is how people get on the 30 under 30 list. As impressive as it sounds, that is paid for.

只是提一下。嗯。

Just saying. Yeah.

不同的预算和策略需要不同的战术。你知道，每个人的方法都不一样。

Different tactics for different budgets and strategies. You know, everyone comes at it differently.

是的，确实如此。好了，我们继续讨论第三个支柱——衡量与追踪。你做了这么多工作。

Yep. It's true. Okay. So let's move on to our third pillar, which is measurement and tracking. So you're doing all this work.

你获得了所有这些引用，在Reddit上与人们互动。你怎么知道这些是否真的有效？

You're getting all these citations out there. You're, you know, inter engaging with people on Reddit. How do you know if this is actually working?

这是最困难的部分之一，因为我认为衡量工具仍在发展中。话虽如此，六个月前还一无所有，完全是一片黑暗。现在我们开始看到一些工具出现，而传统分析根本不够用，它们完全不是一回事。

This is one of the hardest parts because I think the measurement tools are still emerging. That being said, six months ago, there was nothing. It was just a total Now black we're starting to see some of those tools come out. And traditional analytics does not cut it. It's just not the same.

正如我们讨论的，AI流量通常不会带来点击，因此不会显示为典型流量。你可能遇到一个机器人只出现一次，但可能获得数千次提及，而你却看不到任何记录。

AI traffic doesn't often result in clicks, as we talk about, so it doesn't show up as typical traffic. You may get a bot that shows up once, but you may be getting thousands of mentions. And so you don't see any of that.

没错。如果人们因为所有交互都在ChatGPT中进行而不访问你的网站，这些数据就不会出现在传统指标里。

Yeah. And if people aren't visiting your website because everything's taking place in ChatGPT, then that's not going to show up in your traditional metrics.

所以你的谷歌分析流量现在与实际获得的数据可能大不相同。

So your Google Analytics traffic is now very different from what you might actually be getting.

正是。因此你应该关注的一个指标是声量份额——即你在不同平台上的出现频率。现在有些追踪工具可以使用，比如SEO工具SEMRush最近就推出了免费的AI可见度指数工具，能显示品牌在ChatGPT和Google AI模式中的排名。

Exactly. So one of the things that you should look at is share voice. So that's how often you're showing up across different platforms. So there are some tracking tools that you can use. SEMRush, which is an SEO tool, actually just unveiled a free AI visibility index tool, which can reveal how brands rank in ChatGPT and Google AI mode.

这个工具是免费的，值得使用。SEMRush通常表现不错，市面上确实有些可用资源。

That's free. So that's a great tool that you should use. Semrush is generally pretty decent. So there are some things that are out there.

是的。该工具基于对2500多个跨行业提示词的数据分析，是个不错的开端。

Yeah. It's built on data from over 2,500 prompts across a number of industries. A good start.

市面上还有其他一些工具。比如伊桑·史密斯供职的Graphite公司就有一款。虽然选择不多，但确实存在。我认为另一个需要追踪的重要指标是品牌搜索量。如果人们在AI答案中看到你被提及，之后可能会在谷歌上搜索你的品牌名。

So there's a bunch of other tools out there. Graphite, which is the company that Ethan Smith works for, has one. So there are some, but there's not a ton. I think some other metrics that are important to track are branded search volume. So if people see you mentioned in AI Answers, they might Google your brand name later.

这种情况肯定会发生。直接流量。人们看到你被提及后，可能会直接输入你的网址。购买后调查问卷虽然有点老派，但加入这个选项会很有趣。就是那种'您是如何找到我们的？'的调查问卷。

That is definitely something that will happen. Direct traffic. People might type your URL in directly after they see you mentioned. Post purchase surveys, which is a little old school, but is an interesting thing to add in. You know, everyone's had those surveys where they say, how did you find us?

他们会在'您是如何听说我们的'选项中增加谷歌、社交媒体、ChatGPT或AI推荐这些选项，因为这也能为你提供一些指引。

And they'll have, you know, Google or social media, add ChatGPT or AI recommendations to your how did you hear about us options, because that should also give you a little bit of of guidance.

我认为这可能是最重要的事情。现在很多网站都这样做，我也非常迅速地使用这个下拉选项来表明我的信息来源。你甚至可能想进一步细分这些工具，比如是ChatGPT、Gemini还是其他什么。但我觉得你应该现在就行动，在听这期节目的同时，如果你有关于人们如何搜索和找到你的下拉选项，立刻去添加这个选项。

I think that one's probably the most important thing to do. I've seen that now, on a lot of sites, and I'm actually very quick to to use that as the dropdown to let people know that's where I'm coming from. And maybe even you want to dissect that a little more in terms of which of these tools, whether it's ChatGPT or Gemini or something like that. But I think this is something you should do literally right now as you're listening to the episode. If you do have dropdowns for how people search and find you, go add that right away.

好的。这些都很好，但让我们谈谈为什么这个话题在本周变得更加紧迫。我所说的'话题'是指为什么AEO（AI电商优化）在本周变得更加紧迫。9月29日，OpenAI宣布在ChatGPT中推出'Buy It'或'即时结账'功能，允许用户直接在ChatGPT内购买商品而无需离开聊天界面。

Okay. So this is all well and good, but let's talk about why this conversation became even more urgent this week. And I mean, by conversation, I mean, why AEO becomes even more urgent this week. So on September 29, OpenAI announced Buy It in Chatchippy Tea or Instant Checkout. This is a feature that lets people purchase products directly inside ChatGPT without leaving the chat.

这正是我们上周讨论的终极形态。没想到终极形态这么快就来了。

This is exactly what we talked about last week as the end game turns out. The end game is here.

比预期来得稍早一些。

A little sooner than anticipated.

是的。这正说明了这些事物发展得有多快。不是五年后，不是一年后，也不是六个月后，而是现在立刻。

Yeah. And this just shows how quickly this stuff is moving. Not in five years, not in a year, not in six months, literally right now.

没错。这意味着美国ChatGPT用户现在可以直接在聊天中从美国Etsy卖家那里购买商品，很快还将接入超过百万家Shopify商户。这一切将由一个名为'Agentic Commerce Protocol'的系统驱动，该系统基于Stripe构建。这很重要，我们稍后会详细讨论。

Yeah. So this means that US ChatGPT users can now buy directly from US Etsy sellers right in the chat with over a million Shopify merchants coming soon as well. So this is going to be powered by something called Agentic Commerce Protocol, and that is built with Stripe. And that is important, and we'll talk about that in a minute.

比如用户可以进入ChatGPT询问：'我需要为喜欢陶瓷的朋友准备一份生日礼物，预算50美元以内'。ChatGPT会推荐Etsy上的具体商品。点击购买按钮，输入支付信息，就能完成交易，全程无需访问任何网站。这正是我们

So someone can go in and ask ChatGPT, I need a birthday gift for my ceramics loving friend. Keep it under $50 ChatGPT will recommend specific products from Etsy. You tap on the buy button, enter your payment information, and complete the purchase, all without ever visiting that website. So this is exactly what we

之前讨论过的场景。虽然不想说'早就告诉过你'，但事实确实如此。整个发现、考虑和交易过程都在一个对话流中完成。没有重定向广告的机会，销售团队也无法进行后续跟进。

were talking about. Hate to say I told you so, but we told you so. The entire discovery, consideration, and transaction happens in one conversational flow. There's no opportunity for retargeting ads. There's no chance for your sales team to follow-up.

不会有购物车弃置邮件。如果你的产品没被对话提及，你不仅失去了网站访问量，而是直接失去了整笔交易——因为你压根没出现在对话中。

There's no abandoned cart emails. If your product isn't mentioned in the conversation, you don't just lose the website visit. You lose the entire sale because you do not appear in the conversation whatsoever.

从控制角度来看这太疯狂了。你完全被排除在流程之外。

Which is wild from a control perspective. You're completely out of loop.

这彻底颠覆了营销模式。过去二十年的营销方式被完全改变了。

It upends marketing completely. This is the last twenty years of marketing has just been completely shifted.

我认为OpenAI选择的合作伙伴非常具有意义且耐人寻味。他们先与Etsy合作，再选择Shopify的决策既说明问题又对这个群体——对我们所有人来说都极好。这两个平台都围绕中小型企业构建，特别是独立手工艺人、精品品牌和新兴的D2C企业。但重点不在于大型零售商。

And I think who OpenAI partnered with is really significant and interesting. Their deliberate choice to launch with Etsy and then Shopify really is is quite telling or or really great for this group, for all of us. Both of these platforms are built around SMBs, especially independent artisans, boutique brands, emerging D2C businesses. But the important bit is not the larger retailers.

确实。虽然不少大型零售商也与Shopify有深度合作或由其提供支持。但关键是Etsy首当其冲，而且Shopify主要服务大量小型企业而非大公司。我认为这真正实现了我们常说的民主化机遇。这意味着如果你没有谷歌购物的预算或市场优化专长——这正是近年来对Etsy的主要批评，因为要在Etsy上获得排名变得异常困难，必须懂得如何钻营系统规则。

Yeah. And some larger retailers do work with Shopify quite quite a lot or powered by Shopify. But the fact that it was Etsy first, but also the fact that not a lot of large businesses, you know, a lot of smaller businesses work with Shopify as well. I just think it's a really interesting opportunity to democratize which is what we are constantly talking about. And I think it means that if you don't have budget for Google Shopping or an expertise in marketplace optimization, because that's been a real criticism of Etsy over the last couple of years, that it's actually become very difficult to rank on Etsy because you have to know how to game the system.

本质上你不得不为Etsy做SEO优化。这对小企业越来越困难，也是近年部分商家撤离Etsy的原因之一。

You essentially have to do SEO for Etsy. And so that has been increasingly difficult for small businesses. And it's one of the reasons why some small businesses have moved away from Etsy over the last couple Yeah, of

这完美解决了Etsy的最大痛点。从他们的视角很容易理解这一决策。我们称之为对话式电商的模式，彻底改变了传统搜索的运作逻辑。当消费者描述具体需求时，AI代理能根据实际匹配度推荐商品，完全不同于竞价排名的模式。这既创造了新机遇，也带来了新挑战——我们过去构建的整套传统模型突然失效了。

it's a great solve for one of Etsy's biggest challenges. You can definitely see from their angle or their lens why they did it. Conversational commerce, I think is the term we're calling it, really alters this dynamic around the traditional search model. When a consumer describes specific needs, the agent or the AI can surface products based on the actual relevance to those stated criteria. So it's really different from, you know, who paid for prominence, which creates an opportunity, but also, you know, adds to the challenge of how do we show up when we've built an entire model traditionally that is no longer relevant.

我们还得提现在谁在为曝光度付费，因为可以预见这迟早会成为平台的一项功能。

We should also say who paid for prominence right now, because I think we can pretty safely say that that will become a feature at some The point in

经济规律终将胜出

economics will win out

为了

for

当然。

sure.

没错，确实如此。我们已经知道他们正在大量招聘营销和销售团队人员，所以广告投放是板上钉钉的事。

Yeah, for sure. We already know that they're hiring a bunch of people on their marketing and sales team, so ads are coming for sure.

而且技术门槛显著降低，这对普及非常有利。使用Stripe的商户只需更新一行代码就能启用这些支付功能。小企业无需构建定制集成或维护处理这一切的复杂技术基础设施，只需确保产品信息结构化、准确且全面——这点我们之前提过，但现在变得至关重要。

And the technical barrier is remarkably low, which is great for adoption. So merchants processing with Stripe can enable these payments by updating one little line of code. Small businesses don't need to build these custom integrations or maintain sophisticated technical infrastructure that handles all of this. They just need to ensure their product information is structured, accurate, and comprehensive, something we've mentioned before, but now it's critical.

是的。这就把我们带回到AEO（广告效果优化）的话题，因为如果你是小企业或初创公司，现在的问题不是‘我该不该做？’而是‘我能多快为此做好准备？’

Yeah. So this brings us right back to AEO, because if you are a small business or if you're a startup, the question now isn't, should I do this? It's how fast can I get ready for this?

这件事真正重要的意义在于：产品结果是非赞助的自然呈现。它们将完全根据与用户的相关性及用户需求进行排序。回到最初的问题——你的用户或消费者在寻求什么答案？这些内容就会出现在结果中。至少目前OpenAI表示不会为排名收费。

And what makes this really significant, product results are organic and unsponsored. So they'll be ranked purely on relevance to the user and what that user is asking. So going back to our initial, what are the questions that your users or your consumers are asking for? That's what's gonna show up in these results. So at least for now, OpenAI says they're not taking money for placement.

这种情况会改变。但现在你们有个绝佳机会。

That will change. But for now, you've got a great opportunity.

另外需要提到的是，这不只是OpenAI的动作。谷歌也宣布了类似方案叫‘代理支付协议’。整个行业都在朝这个方向发展，OpenAI的公告只是其中最重磅的一个。

So another thing to mention is that this is not just OpenAI. Google announced something similar called agent payment protocol. And so this is the direction that the entire industry is heading. So the OpenAI announcement was the biggest one, but everyone's going to be going in this direction.

我们即将开始使用过多的缩写词。我们已经有了MCG。

We're going to start having way too many acronyms. We've got MCG.

我们明白了。这已经是

We got it. It's already

太多了。APP。我是说，听起来像那首八十年代的歌。你知道吗？不过，可能不是什么大问题。

so many. APP. I mean, it sounds like that eighties song. You know? Anyway, probably not a problem.

所以我带了行话警示器，因为我真的很讨厌行话。

Why I have my jargon buzzer because I really hate jargon.

是的，我同意。但根据SEMrush的研究，到2028年AI搜索流量将超过传统搜索。到2028年？我是说，我觉得到2026年就会实现。

Yeah. I agree. But according to SEMrush's research, AI search traffic is gonna surpass conventional search by 2028. By 2028? I mean, by 2026, I think.

没错。人们已经这样了。我知道很多人讨厌AI模式，但人们已经在用聊天做很多事情了。我上周就说过，我爸用聊天工具选购笔记本电脑。我是说，他没有直接完成购买，但用它来辅助研究。

Yeah. People are already this. I I know a lot of people hate AI mode, but people are already using chat for so much. I I talked about it last week, but my dad shopped for a laptop using chat. I mean, didn't do this purchase process, but he used it for helping him research.

我们就承认吧。坦诚地说，我们都很懒，这确实帮我们省了很多事，所以这趋势不可避免。而且我觉得会比预期来得更快。

Let's just admit it. Let's be open and honest. We're all lazy this really helps us shortcut a lot, so it's gonna happen. Does. It's gonna happen quicker than that, I think.

是的，完全正确。

Yep, absolutely.

好的。让我们快速过一下研究中的几个框架，那些真正能改变你明天能做什么、要做什么以及如何付诸行动的内容。

Okay. Let's hit a few quick frameworks from our research, stuff that actually changes what can do, what you do tomorrow and what you put into action.

首先是内容质量。传统SEO基本上导致大家都在重写相同的文章。我们都见过这种情况，同一个博客文章出现在多个SaaS平台网站上。就像大家都见过的那样。这意味着20个页面中只有一个能带来最多流量或大部分流量。

First, content quality. So traditional SEO basically led to everyone rewriting the same articles. So we've all seen this, the same blog post on multiple SaaS platform websites. Like everyone has seen that. And that means that the result is one out of 20 pages drives the most traffic or most of the traffic.

其余的只是噪音。其余的只是填充物。

The rest are just noise. The rest are just filler.

这对用户体验很不好。说实话，网站体验很糟糕。而AEO（答案引擎优化）则将其推向正确方向，因为最终模型会奖励信息价值。所以需要新颖、具体、真实、可验证的事实。

Which is bad customer experience. Yeah. Websites, quite frankly, are bad experience. Whereas AEO nudges this in the right direction because ultimately the models reward information gang. So new specific, authentic, verifiable facts.

你知道，我们常说统计数据非常重要。如果你的页面听起来像是对相同20件事的重新混音，它就不会被很好地引用。但如果它能真正为每个页面增添有用的内容并清晰地解释概念，那么它就会被引用。

You know, we talk about statistics being so important. If your page sounds like a remix of the same 20 things, it's not gonna get cited well. But if it adds something genuinely useful for each of those pages and explains a concept clearly, then it will.

所以专注于少数能提供真正独特答案的页面。停止把你的精力分散在几十篇看起来和所有竞争对手一模一样的文章上。这已经不再有用了。

So concentrate on a few pages that deliver real unique answers. So stop spreading all your effort across dozens of posts that look exactly like what all your competitors have. That's not gonna be useful anymore.

是的。下一个关键框架是我们所说的多平台现实，这只是个花哨的说法，意思是不同平台的结果差异很大。ChatGPT、Claw、Perplexity、谷歌AI模式等等，平台众多。但结果各异，所以不要只优化一个平台就认为大功告成。

Yeah. The next kind of critical framework is what we call multi platform reality, which is just a fancy way of saying results vary a lot by all of these different platforms. ChatGPT, Claw, Perplexity, Google AI mode. There's a bunch. But the results vary, so don't optimize for one and call it done.

你需要考虑你的结果如何在所有平台上呈现。

You need to think about how your results show up in all of them.

所以要考虑跨多个系统的声量份额。让你的内容在任何地方都能被引用。清晰的声明、来源、结构化事实、人类可读的语言，所有这些都将非常重要。

So think share of voice across multiple systems. So make your content quotable everywhere. So clear claims, sources, structured facts, human readable language, all of that is gonna be really important.

嗯。最后是时机问题。传统SEO中，早期初创公司常常等待。而AEO（AI引擎优化）则需要尽早开始。专注于那些长尾问题，并通过Reddit、维基百科和已发表文章获取可信的第三方引用。

Mhmm. And lastly, timing. So with classical SEO, early stage startups often wait. With AEO, you gotta start early. Focus on those long tail questions and get credible third party citations via Reddit and Wikipedia and published articles.

对于后期公司，继续做SEO中有效的工作。就像我们说的，AEO建立在SEO的基础上，但要在其上叠加AEO。更深入的常见问题解答、产品规格、比较、实施细节、外部引用等等。持续在这些方面积累。

For later stage companies, keep doing what works in SEO. Like we say, AEO stands on the shoulders of SEO, but layer your AEO on top. So deeper FAQs, you know, product specs, comparisons, implementation details, external citations, all of that. Just keep building on all of that.

总之，用更少但更好的页面提供原创信息，为多个AI平台优化，现在就开始你的AEO。好了，我们稍微换个话题，聊聊本周其他一些AI新闻。

So the bottom line, fewer, better pages with original information, you know, optimize for multiple AI platforms, and start your AEO now. Alright. Let's shift gears a little bit and talk about some other AI news from this week.

显然那是最大的新闻。那个故事确实是唯一真正......

Obviously that was the biggest story. That that is that story is the only one that really We

交织了我们的AI新闻。

interwove our AI news.

是的。

We did.

但你知道，还有一些其他我认为与我们的听众相关的好故事。所以

But, you know, there's some other good stories out there that I think are relevant for for our audience. So

关于OpenAI的其他新闻，因为他们这周很忙，他们刚刚发布了Sora二代，这是他们的视频生成模型。但有趣的不只是它能创建视频，而是他们的定位方式。他们构建了一个类似TikTok的浏览界面，可以浏览AI生成的视频，还能进行混剪。

So in other OpenAI news, because, you know, they had a big week, they have just launched Sora two, which is their video generation model. But what's interesting isn't just that it creates video. It's how they're positioning it. They have built a TikTok like feed where you can browse AI generated videos. You can remix them.

你可以分享它们。这与Meta正在做的事情非常相似。他们也在创建一个AI生成的浏览界面。一切都在相互融合。

You can share them. This is very similar to something that, Meta is doing as well. They are also, creating an AI generated feed as well. Everything is merging into each other.

这就是眼球份额的争夺，对吧？OpenAI试图占据你越来越多的时间。但你知道，视频内容制作既昂贵又耗时。所以如果你是初创公司，需要产品演示、解说视频或社交媒体内容，这对你和你的内容制作成本来说可能意义重大。

It's just share of eyeballs. Right? OpenAI is trying to get more and more of your time. But, you know, video content is expensive and time time consuming to produce. So if you're an early stage company that needs product demos, explainer videos, social media content, this could be really big for you and your content production costs.

我要稍微跑题一下，因为我刚听了一个关于好莱坞的播客，特别是关于Sora和知识产权的讨论。显然现在很容易生成知名角色形象，而OpenAI在Dolly时期对此相当严格。但现在他们似乎不那么严格了，因为你完全可以利用一些受版权保护的角色。这将引发一场大战，最终会变得非常有趣，也会带来改变。

I'm gonna go off piste a little bit with this because I was just listening to a podcast about Hollywood and, specifically this podcast was about Sora and IP because apparently it is very easy to generate, well known characters, which is something that that OpenAI was pretty strict about with with Dolly. But I guess they're not so strict about it now because you could definitely use some copyrighted characters. So that is going to be a big fight. That is going to end up being very interesting. That will change.

值得一听马特·贝拉尼关于Sora的那期播客。如果你考虑使用Sora，我觉得可以把这点记在心里。

It's worth a listen to that podcast on Sora with Matt Bellany. If you're thinking about using Sora, just bear that in mind, I think.

是的。最后我想说的是，那个TikTok风格的推荐结果，显然传播性很强。这是个很棒的发现机制，人们能找到并混剪内容，这意味着你AI生成的视频有可能自然触达大量受众。

Yeah. Lastly, I'll just say this about, you know, the result, the TikTok style feed, you know, it's obviously very viral. It's great discovery mechanism. So people can find and remix content, which means your AI generated videos could potent a lot of, reach high level of audiences organically.

关于这点我想说，前几天我研究过Meta的那个功能。不知道人们是否真的需要这个。大家对Meta的AI垃圾内容流的反应并不积极，他们遭到了很多反对。所以问题在于，人们真的想要这个吗？

I think the one thing I will say about this, and I was doing some research on the meta one the other day. I don't know if people actually want this. The the reaction to the meta AI slop feed was not positive. Like, they got a lot of pushback. So I think the question is, do people want this?

是啊。问号。让我们拭目以待。

Yeah. Question mark. Let's see.

其实'AI垃圾'这个词昨晚我和朋友聊天时也提到了。杰斯，或许我们该快速定义一下这个词，因为它会反复出现。

You know, that term AI Slop actually came up, in one of my conversations last night with a friend. So Jess, maybe maybe we should define that real fast and then because it will keep coming up, I think.

没错。所谓AI垃圾本质上就是现在TikTok上那些劣质AI视频——虽然已经存在了——比如你的狗在天上飞之类的，你知道的，就是那种愚蠢的垃圾内容。我们准备做一期关于AI垃圾的节目，因为这问题会越来越严重。

Yeah. So AI slop is essentially all those shitty AI videos that you see on TikTok now, actually, they do already exist, but it's things where like your dog flies across the sky or something, you know, stupid, stupid shit like that. We're gonna do an episode on AI Swap because it's gonna increasingly become a really big problem.

但我觉得它还包括那些质量低劣的内容，就是那些非常...嗯...容易...

But I think it also encompasses like the poor content that you see that's very Yeah. Easily

是的，百分百确定。因为最终AI会像一条咬着自己尾巴的蛇，或者说‘衔尾蛇’现象，它会在自己的输出上不断训练。然后就陷入了糟糕的循环。没错。

Yeah. A 100%. Because eventually the AI is going to be like a snake eating its own tail, an aurora boros, if you will, because it's going to train on that. And then it's just training on shitty cycle. Yeah.

对，正是如此。不过我们没时间深入讨论了，还有一堆其他内容要讲。

Yeah. Exactly. But anyway, we don't have time to get into that. We've got a bunch more stories to cover.

没错。我们来说说下个重点——这对使用Slack的人（可能包括我们大多数人）很重要。Anthropic刚刚为Slack工作区推出了原生Claude集成，现在你可以直接在Slack频道里使用Claude了。

Yeah. Let's talk about yeah. This next one's really big for anyone using Slack, which is probably a lot of us. Anthropic just launched a native Claude integration for Slack workspaces. So you can now have Claude directly in your Slack channels.

这不仅仅是聊天机器人。它实际上能访问你的Slack历史记录，理解对话上下文，真正提供帮助，而不是像那些被我们忽略的通知——大家都知道Slack里总有被无视的通知。

And this isn't just a chatbot. It can actually access your Slack history. It can understand context from your conversations, and it can actually be really helpful instead of just another notification that you ignore because I think we all have those notifications that we ignore on Slack.

是的，这个功能让我很兴奋。你可以用Claude总结超长讨论串或起草回复。有些人把客服对话接入Slack，这样就能让Claude查看并发现问题。但说到底，我现在经常找不到需要的内容，这个功能应该能解决这个问题。

Yeah. I'm pretty excited about this one. You can use Claude to summarize super long threads or draft responses. Some people have their customer support conversations feed into Slack, so you could actually have Claude take a look at that and identify issues. But just ultimately, I'm getting to the point where it's I'm having a really hard time finding things, and I think this will help solve that problem.

完全同意。我上家公司几乎不用邮件，所有事都在Slack上完成。但Slack的搜索功能出了名的难用——当你确信看过某条消息却死活找不到时，因为频道实在太多了。现在它和邮件一样糟糕。不过我觉得最实用的场景是会议纪要。

Absolutely. My previous company barely used emails. They they did almost everything on Slack, and Slack is notoriously difficult to search for things like, you know, when you're absolutely convinced that you saw something and you cannot find it for the life of you because there are too many channels. It's actually just as bad as email at this point. But I think the practical use case that I keep thinking about are meeting notes.

如果你有为项目创建的Slack频道，Claude可以回顾对话内容，生成决策摘要、行动项和待解决问题。这对没有专职项目经理的小团队特别有用。不过隐私问题值得考虑，要清楚Claude能访问哪些数据，合理配置权限。毕竟不是所有Slack内容都该让AI读取。

So if you have a Slack channel for a project, Claude can review the conversations and generate a summary of decisions made, action items, open questions. Like, I think that is genuinely useful for small teams that don't necessarily have a dedicated project manager. Think the privacy considerations are worth thinking about, though. Make sure that you understand what data Cloud has too and configure permissions appropriately. You know, not everything in your Slack should be AI readable.

一如既往，这是需要牢记的一点。

Just something to bear in mind as always.

好的。现在是大家最爱的环节——‘AI失控现场’。本周的故事尤为应景，因为它展示了当AI内容生成不受约束时会发生什么。

All right. Time for everyone's favorite segment, AI Gone Wrong. And this week's story is particularly relevant because it shows what happens when AI content generation goes unchecked.

这其实和我们刚才讨论的AI惰性有点关系。Spotify去年删除了7500万条垃圾音轨。7500万，这个数字太荒谬了。

This is actually a little bit of the AI sloth that we were talking about moments ago. So Spotify just removed 75,000,000 spam tracks in the past year. 75,000,000. That number is ridiculous.

造成这种爆炸式增长的原因显然要归咎于AI。AI生成的音乐让创建和上传数千首曲目变得异常简单。这不是艺术家创造性使用AI工具的辅助创作，而是纯粹为了钻流媒体版税系统空子的全自动低质量曲目。

And the reason for this explosion, obviously, the blame sits with AI. So AI generated music has made it really easy to create and upload thousands of tracks. So this is not AI assisted production where artists use AI tools creatively. We're talking about fully automated, low quality tracks designed purely to game the streaming royalty system.

骗局是这样运作的：有人用AI生成数百或数千首通用曲目，可能是氛围音乐、低保真节拍、白噪音等等。他们用数十个假艺术家名义上传到Spotify，再通过机器人或点击农场人为刷高播放量。

So here's how the scam works. Someone uses AI to generate hundreds or thousands of generic tracks. So that could be ambient music, lo fi beats, white noise, whatever. They upload them to Spotify under dozens of fake artist names. They use bots or click farms to artificially inflate play counts.

每次播放都会产生微薄的版税。但如果用数千首曲目的数百万次虚假播放来累计，这就变成了实实在在的钱。

Each play generates a tiny royalty. But if you multiply that by millions of fake plays across thousands of tracks, and that actually becomes legitimate money.

没错。我对此的不满倒不一定是这些曲目的创作方式，而是他们用机器人农场伪造听众来牟利。这才是真正的问题所在——不在于内容是AI生成的，而在于他们用这些AI机器人刷高播放量来获取更多收益。

Yeah. And and my issue with this is not that necessarily the the tracks were created. It's that they use the bot farms to, you know, fake people listening to it to then create money. That's that's the real problem. Not that not that it was AI generated content, but that they use these AI bots to bump up the price so that or bump up the listen so they would get more money.

因为归根结底，仔细想想，Spotify获得了一个资金池，然后根据整体池子的情况分配给艺术家。所以如果其中一部分被这些AI机器人农场抽走，是的，从Spotify的总资金中，你实际上是在伤害艺术家。

Because ultimately, if you think about it right, Spotify is gets a pool of money, and then they distribute that to artists based on the overall pool. So if a section of that is being taken away for these AI bot farms Yeah. From the overall money that Spotify has, you're really harming the artists.

是的，完全正确。

Yeah. Absolutely.

而且，你知道，这并非理论上的。我是说，Spotify估计500万首曲目是个很大的数字，这些曲目在被打击前通过欺诈版税产生了数百万收入。所以，这又是从合法艺术家那里窃取的大量资金。

And and, you know, this isn't theoretical. I mean, Spotify estimates 5,000,000 tracks is a lot, and those tracks were generating millions in fraudulent royalties before they were cracked down. So, again, that's a lot of money being stolen from legitimate artists.

我觉得特别阴险的是——有个词可以形容——其中一些曲目被设计成出现在算法生成的播放列表中，比如Spotify DJ。所以真实用户最终会在不知情的情况下收听这些AI垃圾内容，这确实污染了整个生态系统。

I think what's particularly insidious, there's a word, is that some of these tracks are designed to appear in algorithm generated playlists, so Spotify DJ. So real users ended up listening to them without realizing that they're AI spam. So it does really pollute the entire ecosystem.

现在我觉得特别讽刺的是，为了应对这个问题，Spotify实施了新的AI政策和所谓的垃圾过滤强制执行。所以他们正在用AI来检测AI生成的内容。所以这个

Now here's the bit that I find particularly ironic. In order to combat this, Spotify implemented new AI policies and what they're calling spam filter enforcement. So they're using AI to detect AI generated stuff. So the

检测机制主要关注几个方面：异常上传模式（比如一周内一个账号上传500首曲目）、相似性分析以识别明显是同一AI生成模板变体的曲目，以及看起来像机器人活动而非真实用户的互动模式。虽然我可能也会被算作机器人活动，因为我反复听同样的歌，尽管它们绝对不是AI生成的。

detection, it focuses on a few things. So unusual upload patterns like one account uploading 500 tracks in a week, similarity analysis to identify tracks that are clearly variations of the same basically AI generated template and engagement patterns that look like bot activity rather than realism. I don't know. I can be counted as bot activity because I listen to the same songs over and over and over again, although they are definitely not AI generated.

但我觉得，也许那是正常行为，因为感觉我们都有这种习惯。不过，我是说，它听着

But I think that I mean, maybe that's normal behavior because I feel like we all have that. But, I mean, it listen

听听我们的说唱，看看年底时我们的说唱会带来什么。

to our rap there, see what our rap says at the end of the year.

没错，没错。我觉得最有趣甚至滑稽的是，最终这演变成了双方之间的军备竞赛。随着Spotify更擅长检测AI垃圾信息，那些垃圾信息制造者也会更擅长让AI内容看起来更真实。于是就成了AI对抗AI的局面。

That's right. That's right. Where I think this is fascinating and maybe hilarious is that, ultimately, you get this arms race between the two. So as Spotify gets better detecting AI spam, the spammers are gonna get better at making their AI content look more legitimate. So we just have AI versus AI.

而且我不觉得...那是个奇怪的'天网'现实。我不确定...具体会走向哪里

And I'm not that's a that's a weird Skynet reality. I'm not sure Sure where that

那么，在这个例子中，我们的观众应该从中吸取什么教训呢？

is. But what what what is the takeaway for our audience in this instance?

归根结底还是要警惕AI生成内容。如果你的策略是让AI生成数百条内容，然后不加甄别地充斥网络，缺乏真实视角和可验证的真实性，你终将被发现。平台正在提升检测能力，一旦被发现，你就会遭到封禁。

Always goes back to being a cautionary tale about AI generated content. So if your strategy is I'll just have AI generate hundreds of pieces of content and flood the internet with it without, like, a real perspective, without some of that on test authenticity, you're gonna get caught. Platforms are getting better at detecting this, and when they do, you're gonna get banned.

我认为这一切都回到了我们一直在讨论的AEO（真实参与优化）——真实性和质量至关重要，因为这才是人们真正想要的。即使使用AI，你也必须保持真实的语调。这仍然是绝对关键的因素，而且平台的检测能力只会越来越强。因此人类的洞察力和参与度反而变得更加重要。

I think all of this just comes back to what we keep talking about with AEO, is that authenticity and quality matter, because that is what people actually want. So even if you're using AI, you should absolutely still have that authentic tone to it. There still has it's that is still absolutely key, and the platforms are only gonna get better at detecting it. So that is where human insight and human involvement becomes even more important.

是的。如果你在内容创作中使用AI（这当然应该用，我们显然鼓励这样做），但要确保有真实的人工监督和编辑，确保你在增值、有独特视角。AI是帮助你创作更好内容的工具，或帮你克服'空白页'的挑战，但它不能替代真正的内容创作，更不能用来生产大量垃圾内容。好了，杰斯，我们今天就到这里结束吧。

Yeah. So if you're using AI in your content creation, which you absolutely should, obviously, we we encourage that, but make sure there's genuine human oversight, editing, that you're adding value, that you have a perspective. The AI is a tool to help you create better content or to help you with that challenge of the blank page, but it's not a replacement for actually just creating content and producing a ton of slop. Okay. So, Jess, we're gonna wrap up here.

我们提供了大量信息。哦，一如既往。一如既往。但我们会尽量从这部分内容中给你一些非常清晰的可执行事项。我们希望让这些建议切实可行。

We've given a lot of information. Oh, as always. As always. But we're gonna try to give you some really clear action items from from this bit. We want to make this practical.

那么关键行动项第一条——我们经常讨论这个——进行可见性审计。访问SEMRush AI可见性指数，在ChatuchuBT、Perplexity、谷歌AI模式下搜索你的品牌，查看显示和未显示的内容，然后截图或录屏。这将建立你的基准线。

So the key action item number one, we we talk about this a lot, but do a visibility audit. Go to the SEMRush AI Visibility Index, search for your brand on ChatuchuBT, Perplexity, Google AI mode, see what shows up, what doesn't, and then take screenshots that or record it. This is gonna create your baseline.

我们推荐SEMRush工具是因为它目前免费。趁还能用的时候好好利用。行动项第二条是之前讨论过的问题挖掘。花几个小时整理所有销售通话记录、客服工单、行业内的Reddit讨论串。这些都可以导入ChatGPT帮你识别，但最终要整理出50到100个客户实际提出的问题清单。

And we talk about the SEMRush tool because it's free for now. So take advantage of that while you can. Action item number two is this question mining that we talked about earlier. So take a couple hours, go through all your sales call notes, go through your support tickets, your Reddit threads in your industry. All of that can be dumped into ChatGPT and it can help you identify all of these, but create a list of 50 to 100 actual questions that your customers are asking.

所有内容都会包含其中，但这将帮助你朝正确方向前进。是的。

And it will all be in there, but that is going to help guide you in the right direction. Yeah.

查看你的客服系统，筛选出高频问题并逐一解答。行动项第三条，考虑做出一次真实贡献。找一个Reddit讨论串、论坛帖子或在线讨论，针对你业务能解决的问题真诚提供帮助。表明身份后分享有用信息，注意不要像垃圾广告。先尝试一次，感受效果，观察互动情况，看看如何帮助你改进现有工作。

Take a look at your customer support system and pull out the top questions and then go through and answer those. Action item number three, think about one authentic contribution. So find a Reddit thread or a forum post or an online discussion where you can genuinely help and provide some information because it's something that your business solves. Identify yourself, share some useful information, you know, that won't be spammy. Just and try one, see how it feels, see what the engagement, and see how that, you know, basically helps you improve, what you're doing.

第四条，测试你的结构化数据。选择最重要的产品或服务页面，让ChatGPT进行总结。信息是否清晰准确？如有缺失或混淆之处，立即修正。

Item number four, test your structured data. So pick your most important product or service page, ask ChatGPT to summarize it. Is the information clear and accurate? If not, what's missing or confusing? Fix that.

所以Jess，要测试、测量，就像我们常说的：持续迭代永无止境。

So Jess, test, measure, and as we always talk about, always be iterating.

持续迭代。

Always be iterating.

始终如一

Always be

这将有助于你的AEO（广告效果优化）。

gonna that's gonna help you with AEO.

就这样吧。我们结束了。

With that So Oh. We're wrapped up.

本周的内容就到这里。和往常一样，我们向你提供了大量信息，希望你觉得既实用又有趣。记住，你可以在所有社交媒体平台上找到我们——TikTok、Instagram、YouTube、LinkedIn，所有你能想到的地方。我们无处不在。

That is that is us done for the week. As always, it's a lot of information that we are throwing at you, but we hope that you find it useful and interesting. And remember, you can find us on all your social media channels. We are on TikTok, Instagram, YouTube, LinkedIn, everywhere you I don't know. Everywhere you can find anything, we are there.

我们的账号是early_adopter（注意adopter没有字母e）。

We are at early underscore adopter. Remember, there is no e in adopter. Make

记得在所有平台点赞、订阅、关注，该做的操作一个都别落下。

sure you like, subscribe, follow, whatever you're supposed to do on all those channels, make sure you do that.

没错，确实如此。这真的能帮助我们提高曝光度。说到这里，我们期待下周与您详谈。祝您一周愉快，再见。

Yeah, exactly. It does really help us, be discoverable. And on that note, we look forward to talking to you next week. Have a great week. Bye.

谢谢，杰西。

Thanks Jess.