本集简介
双语字幕
仅展示文本字幕，不包含中文音频；想边听边看，请使用 Bayt 播客 App。
举个例子，我们的软件设计上就限制了AI每天联系房主不超过三次。
For example, what our software does, it actually, by design, prohibits from our AI contacting a homeowner more than three times a day.
对吧？
Right?
我们这么做是有道理的，因为如果你用同一个号码一天内联系某人超过三次，像AT&T、Verizon和T-Mobile这些运营商很可能在几分钟内就把你的号码标记为垃圾号码。
So we do that for logical reasons because if you contact someone, and this is like a general point, more than three times a day using the same phone number, AT and T, Verizon, and T Mobile will probably mark your number as spam within like couple of moments.
对吧？
Right?
所以我们还开发了一套自己的号码健康监测系统。
And so something we do is we have also our own like number health monitoring software.
我们学到的一个重要经验就是，一天别给同一个人打超过三次电话。
And one of the lessons we learned is to not call people more than three times a day.
这一点已经直接写进了我们的系统里。
So this is something that's built into our system.
对吧？
Right?
如果你想要同一个客户联系四次，我们的系统会直接阻止拨出电话。
If you want to contact, like, the same customer four times, we will actually, like, block the phone call.
对吧？
Right?
我们正是通过这种方式，防止AI被滥用。
And this is how we're trying to help the AI being misused.
别把AI看成是完全取代人类的东西。
Think about AI not as something that totally replaces people.
客户其实很欢迎这种安排，因为他们再也不用做那些烦人的工作了，而这些正好是AI能胜任的。
The the human qualities are happy to have on their end because they don't have to do, like, annoying stuff anymore, and it's something that that AI can do.
今天的节目由我们的合作伙伴之一赞助播出。
Today's episode is brought to you by one of our
赞助商是UpFrog。
show partners, UpFrog.
说到线上价格透明，UpFrog可是开山鼻祖。
UpFrog is the OG when it comes to pricing transparency online.
他们通过付费广告找到系统更换的潜在客户，并建立了一整套流程，用来挖掘这些难得的系统更换机会、培育线索，并在你的团队介入之前就促成成交，帮你拓展业务。
They find system replacement leads through paid ads and have developed an entire system to drive those hard to find system replacement opportunities, nurture the leads, and book them into sold systems before your team arrives to help you grow your business.
你好。
Hey.
我是MarketStorm的Christa，我们为全国各地的家政服务公司提供支持。
I'm Christa with MarketStorm, and we work with home service companies across the country.
我有个问题。
Quick question.
如果你们当地另一家暖通或水管公司，在客户还没上谷歌搜索之前就已经影响了他们，那第一个接到电话的会是谁？
If another HVAC or plumbing company in your area starts influencing homeowners before they ever hit Google, who gets the first call?
等客户搜‘附近水管工’的时候，他们其实早就开始研究问题了。
By the time someone searches plumber near me, they've already been researching the problem.
我们用AI来捕捉这些早期的购买意向信号，让你的品牌提前出现在客户面前。
We use AI to detect those early intent signals and put your brand in front
在他们面前。
of them.
所以当他们搜索时，你的名字听起来已经很熟悉了。
So when they do search, your name already feels familiar.
访问 marketstorm.ai 或者给我们发短信到 (213) 575-5448。
Visit marketstorm.ai or text us at (213) 575-5448.
大家好。
Hello, everyone.
我是这个播客世界的主持人。
I'm there in podcast world.
希望你今天过得特别好。
Hope you're having a wonderful day.
你正在收听或观看《服务企业精通》播客。
You're listening to or watching Service Business Mastery podcast.
我是主持人之一特什·布利塞特，现在虚拟地坐在我联合主持人约书亚·克劳奇旁边。
I am one of your host, Tersh Blissett, sitting virtually next to my cohost, Joshua Crouch.
今天我们请到了凯文，来自 pingai.com。
And we have Kevin on the show today with pingai.com.
今天我们来聊聊AI语音，还有AI在维修行业里的优缺点。
And we're gonna talk about AI voice, good, bad, and ugly about just AI in the trades.
如果你是第一次听我们的节目，欢迎你！
And so if you're new to the show, welcome.
我们经常讨论维修行业里的AI和自动化问题。
And we talk a lot about AI and automation in the trades.
如果你已经关注我们很久了，肯定知道这一点。
If you've been here for a while, you know that.
我们都很期待。
We're excited.
如果你正在直播观看本期节目，随时欢迎提问。
If you're live streaming this episode, plea please feel free free to ask any questions.
我突然有点舌头打结。
My tongue is tied all of a sudden.
不过，我真的很兴奋能和Kevin聊聊，听听他对这个行业未来走向的看法。
But, yeah, I am super excited to talk to Kevin and get his take on where things are going in our industry.
那么，欢迎你来到节目，凯文。
With that being said, welcome to the show, Kevin.
非常感谢。
Thank you so much.
当然。
Absolutely.
乔什，你有什么要补充的吗？
Josh, what you gotta add?
我知道你早就迫不及待了。
I know you're chomping at the bit here.
是的。
Yeah.
我们要聊的是线索响应速度，这可是我经常提到的话题。
I mean, we're gonna talk about speed to lead, and this is, you know, something that, I know I post about a lot.
我知道特什，还有PhoneTAP的另一位联合创始人，也经常分享这个话题。
I know Tersh and, one of our cofounders of PhoneTAP has also been posting about a lot.
这其实是呼叫中心和家装、家政服务类企业的普遍难题，说到底，几乎任何行业都一样。
It's just the the massive problem of call centers and home improvement, home service businesses, any it's really any business.
甚至都不用非得是我们这个领域。
It doesn't even matter if it's in our industry.
这是所有企业的通病。
It's every business.
我自家后院最近搞了个长达十八个月的大工程， landscaping 和各种装修都搞了一遍。
Like, I've had a massive eighteen month project going on in my backyard and stuff with landscaping and everything else.
想联系上这些承包商，问个准话，让他们接电话、回消息，简直让人崩溃。
And getting a hold of these contractors and getting firm deadlines and getting them to answer the phone and getting them to get back to us is literally a nightmare.
真的，
Like,
就像你指着一个地方说，
it's like pointing to
过去这十八个月里我能讲出无数类似的故事，我相信每个人都有过这样的经历。
stories I could tell over the last eighteen months are absolutely insane, and I'm sure everyone's got stories like this.
我见过有人联系四个人，结果只有一个人回电。
They I've seen people, they they call four people and, like, one gets back to them.
你花那么多钱做营销，让人在需要你的时候能找到你，可到了最后关头却掉链子。
And I'm like, you spend so much money on marketing and getting your business to show up when someone needs you, and then you just fumble the ball at the goal line.
没错。
Like Yeah.
我最贴切的比喻就是，你被抄截了。
That's the best analogy I can use is, like, you know, you get an interception.
你都冲到端区了。
You get to the goal.
这种精彩片段大家肯定都看过。
You you've all seen those highlights.
明明冲到底线了，结果却把球掉出界了。
You get to the goal line and, like, you fumble it out of bounds.
简直太可惜了，可这种事每天都在各行各业上演。
You're like, it's just a waste, but it happens every single day in all these businesses.
所以我特别期待聊聊凯文正在做的事。
So I'm super excited to talk about what Kevin's got going on.
凯文，能不能简单介绍一下你自己？让还不认识你的人了解一下你是谁、做什么的。
Kevin, could you just do an introduction on yourself and let for people that don't know who you are, but who you are and what you're
我是来自Leaping AI的凯文。
I I am Kevin from from Leaping AI.
我们为服务行业提供AI语音和AI短信解决方案。
We do AI voice and AI texting solutions for the trades.
我出生在德国，是在德国长大的。
I am originally from Germany, grew up in Germany.
我学的是数学和计算机科学。
I studied mathematics computer science.
毕业后，我本来想直接创业，但发现自己缺乏做生意所需的技能。
After that, I really wanted to already found something after college, but decided, hey, I I actually lacked the business skills to become an entrepreneur.
现在我才明白，其实只要敢闯，就能当创业者。
And well, now I know that one could just go for it, right, and become an entrepreneur.
但那时候我觉得得先接受一些商业教育，于是就加入了一家叫波士顿咨询公司的咨询企业，待了三年，给首席执行官们提供建议，从飞机制造选址到半导体行业的数字战略，什么都涉及。
But back then, I thought I had to get some kind of business education, so I actually joined a consulting company called The Boston Consulting Group, and worked there for three years advising CEOs on on everything from where to manufacture airplanes to digital strategies for semiconductors.
不错。
Nice.
这么干了三年后，我觉得行了，现在该把商业技能都攒齐了。
Did that for three years, and then at some point I was like, okay, now I have all the business skills.
我早年学的技术底子还在。
I already have the tech skills from back then.
那就再试一次创业吧。
Let me try again at at at being an entrepreneur.
大概三年前我创办了这家公司，最初是在欧洲。
Started this company around three years ago, at first in Europe.
可惜欧洲的创业环境跟美国比起来差远了。
Unfortunately, the entrepreneurship scene in Europe is way smaller than, like, in The States.
那儿的人很少选择创业，所以我找投资特别难。
People don't really become entrepreneurs, and so it was really hard for me to get funding.
我当时特别艰难，全靠自己撑着，把从前工作攒下的钱都花光了，甚至用客户付的钱来发下个月的工资。
I was struggling a lot, bootstrapping, used up all the resource I had from my previous job, essentially using money that customers paid us to cover salaries for for the next month.
那时候我就下定决心，得来美国了。
And that's when I decided, you know, I gotta gotta come to The US.
对吧？
Right?
只有在美国，才能拿到投资。
That's where you get that's where you get funding.
我们后来确实
And We got
在这儿拿到了不少资金来发展。
lot of money to grow around in this country.
是啊。
Yeah.
我不清楚钱都从哪儿来的，但确实有。
I don't know where it all comes from, but, like Yeah.
这些估值，我不太确定是Claude还是Anthropic
These valuations of the like, I I don't know if it's Claude or Anthropic
还是Anthropic。
or Anthropic.
是的。
Yeah.
又融了一大笔钱，多得离谱，超过了大多数国家的国内生产总值。
Raised another, like, obscene, like, obscene amount like, more than the GDP of, like, most countries.
这太夸张了。
That's wild.
就一家公司的估值。
Evaluation of one company.
简直不可思议。
It's absolutely insane.
对。
Yeah.
现在有好多投资者呢。
Like, there's a lot of investors.
五年前根本还没这回事。
That didn't, like, exist, like, five years ago.
对吧？
Right?
或者差不多是这样。
Or something like that.
所以这就意味着
So That's
最让人不可思议的就是这一点。
the that's the craziest thing about it.
是啊。
Yeah.
我一年半前来到美国，之后事情就一件接一件地发生了。
So I I came to The States a year and a half ago, and then it was one thing led to another.
我参加过旧金山的一个创业加速器，叫Y Combinator。
Like, I did a company accelerator called Y Combinator in in San Francisco.
参加了那个项目后，融资变得容易多了。
Doing that, it becomes very easy to raise funding.
所以我们确实融到资了，没倒闭。
So then we we actually did raise funding, didn't go bankrupt.
对吧？
Right?
也不用再自己掏钱硬撑了。
Didn't have to bootstrap anymore.
那是一年半前的事了，从那以后我就一直住在旧金山，努力把公司做大，嗯，说起来简单，其实也没那么简单。
And so this was a year and a half ago, and ever since then, I've been living in San Francisco, trying to grow this company, and and yeah, that's that's the short well, wasn't that short.
这是我人生经历的一个简要版本。
The the kind of short version of of my life.
是的。
Yeah.
不，这太棒了。
No, it's awesome.
说来有趣，你提到你觉得自己得先接受商业教育。
You know what's funny, you mentioned like the you thought you had to get a business education.
虽然这跟我们现在聊的没关系，但你是在创业前接受商业教育的吧。
I know it's not what we're talking about, but business education before you started a company.
我很庆幸自己创业前干过技术活，那段时间学到了很多，对后来做数字营销公司帮助很大。
And I'm grateful that I was in the trades beforehand because I learned a lot of lessons that helped me with my digital marketing company.
但话说回来，这事儿真挺有意思的。
But even still, it's funny.
你总会经历起起落落，以为自己已经懂了，以为万事俱备，可一到下一阶段，才发现还得学新东西，根本没准备好。
Like, you go through these peaks and valleys where you think you you think you got it, you think you know it, and you get to the, like, the next level, and you're like, oh, I gotta learn some new skills because I'm not ready for this.
说到底，根本不存在什么完美的时机。
And it's like, it's it's like, there's just no perfect moment.
我知道很多听众是技术人员或者公司经理，他们也想自己创业，所以才会听这个节目。
And I know a lot of people listen this because we have technicians or general managers of companies that wanna start their own business that listen to this show too.
我觉得我们总是等着那个完美的时机。
And I think a lot of times we're always waiting for that, like, perfect moment.
但根本就不存在什么完美的时机去做这件事。
But there there is no such thing as a perfect moment to do this.
你得自己动手去做。
It's you you you do it.
有时候，最精彩的故事恰恰发生在你身无分文、靠自己硬撑的时候，就像你提到的那样，因为你别无选择。
Sometimes the best the best stories, even though it's the most stressful, are when you don't have any money and you're bootstrapping like you mentioned, because you have no choice.
你已经被逼到墙角，只能上阵了，必须想办法闯过去。
Your back is up against the proverbial wall, and it is go time, and you gotta figure it out.
有时候，最好的结果恰恰就在这时候发生。
And sometimes the best things happen.
有时候你会失败，但你会学到下次不该怎么做。
Sometimes you fail, but then you learn what not to do next time.
在我的工作中，我经常接触很多这类手艺类企业的老板，有些公司规模高达五千万、六千万甚至七千万美元。
And In my in my in my line of work, I deal with, like, a lot of CEOs of of these trades businesses, sometimes, like, $50.60, $70,000,000 businesses.
我不确定这是不是主观感受，但我真没遇到过哪个CEO是上过大学或大学毕业的。
I'm not sure if this is like just a subjective feel, but I don't think I've met a single CEO who actually went to college or graduated college.
大多数人高中毕业就进了这个行业，二十年后就成了百万富翁。
So most people, they did high school, and then they you know, got into this industry, and now twenty years later, they're like millionaires.
我都在想，我这一生到底哪里做错了，明明在科技圈里拼命打拼，结果呢。
I'm like, what did I do wrong in my life, you know, that I'm grinding the tech scene, and yeah.
我当初为什么要拿个学士学位？
Why why did I even have a bachelor's degree?
我学那些东西到底图什么？
Why did I study anything, right?
我还不如直接开个空调安装公司或者换窗户的生意呢。
When I could have just like started my own like HVAC business or window replacement company.
是啊。
Yeah.
你这么说真有意思，我记得我2003年高中毕业时，整个就业市场都在说，你必须上大学。
It's really funny you mentioned that because I remember I graduated high school in o three, and at at that point, like, the whole job market and stuff was all about like, you gotta go to college.
你非得拿个学位不可。
You gotta get a degree.
你得搞一大堆这些东西。
You you gotta, you know, have all this stuff.
所以我回去上学，花了不少钱，还背了一身债。
And so I went back and did school, and I spent a lot of money and got a lot of loans.
而且我
And I
那对你来说结果怎么样？
just How'd that work out for you?
我一样都没用上
I I didn't use any of
它。
it.
就像
Like,
我知道在美国这很普遍，而且我觉得确实已经变了好多。
I know that's a common thing here in The States, and I think that's I I do think that has changed quite a bit.
比如，特什是不是上常春藤盟校了？
Like, Tersh has got a is Ivy 17?
18岁？
18?
对。
Yes.
他今年17岁。
He is 17.
他今年十月就满18了。
He'll be eight in October.
是啊。
Yeah.
快18岁了。
So almost 18 year old.
我有个18岁的孩子，另一个儿子也是18岁，他今年七月就满18了。他们俩都在考虑技工类的职业，比如电动车、汽车维修、暖通空调，或者焊接之类的。
I have one 18 year old, another eight my son, 18 year old, is gonna be he's gonna be 18 in July, and they're all looking at trades careers or some variation of the trades, electrical cars, auto, HVAC, whatever, welding.
他们能选择的路特别多，看到这一点真的挺让人欣慰的。
You know, there's all they've different options open to them, which is really cool to see.
我们还做了一些相关节目。
And we do some some episodes.
最近刚做了一期，讲的是技工行业怎么招聘、怎么晋升这些事，怎么操作。
We just did one recently about, you know, hiring and and promoting and stuff in the trades and how to do that.
但这个领域挺有意思的，一方面吸引了不少年轻人想进技工行当，另一方面又有大量科技资金涌入，用来开发高级自动化系统，帮这些公司提升效率——毕竟这行一直都很依赖人力。
But it's an interesting mix because we're getting people the young people that wanna come into the trades, but then we're also getting a lot of a lot of tech money, tech support to build advanced automations and make these companies more efficient at least as from because it's always been a very labor enhance labor intensive business.
做这些生意，光是后台办公支持就得配不少人，才能运转起来。
You need a lot of people, a lot of support on the back end in the office to make these businesses run.
但通常来说，老板们想多赚钱，就会尽量压缩后台人员，结果就带来了我们今天要聊的问题：客户电话总是接不及时。
And then typically, the way that these guys make more profit is they have less support on the back end, which causes the problems we're gonna talk about today where we don't answer customers fast enough.
电话总是没人接。
We don't answer the phone fast enough.
我们连短信都不回。
We don't text back.
表单提交了不及时回复，晚上或周末也没人理，这问题可不小。
We don't answer form submissions fast enough, or we don't answer after hours or weekends, and it causes a big problem.
这里面浪费的东西太多了。
And there's a lot of waste there.
我们在和老板们聊天时发现，很多人根本没意识到问题有多严重。
And some of that lot of that waste we're finding when we have conversations with business owners is they don't even they don't even know the gravity of the problem.
一旦他们看到潜在损失的金额，立马就紧张了。
They and once they see the number, like the dollar amount that's potentially lost, it gets their attention really quickly.
所以，凯文，你当初决定创业时，为什么选这个方向？
So why you know, Kevin, when you you decide to start a business, why why this?
你看到了什么问题，才决定创办这家公司，专门帮家装和家政公司用AI代理解决问题？
What what problem did you see that wanted you to create this sort of company that helps with AI agents for home home improvement and home service companies?
我当初想创业，是因为在大学当实习生时，曾在亚马逊的呼叫中心工作，那时候我就意识到，这是一门职业。
So the original impulse to start this business was when I worked in the call center at Amazon as an intern back in college, and I saw that this is a profession.
对吧？
Right?
这跟实际的业务问题脱节了。
So this is detached from the business problem.
更像是陷在细枝末节里。
It's more like being in the weeds.
对吧？
Right?
这工作其实挺没意思的，人员流动特别大，我跟承包商聊的时候，他们也经常遇到这问题。
This is a profession that's not really fun, where you have super high fluctuation, and this is also when I talk with a contract, it's also something that they deal with.
对吧？
Right?
就是人员流动大。
Like fluctuation.
这工作不像call center的客服，没人会干上四十年。
It's not a job that like a call center agent, it's not something that people wanna do for forty years.
这其实就是我当初的初衷。
And so this was my original impulse.
我那时候更多是从底层出发的。
I was coming more from, let's say, the bottom up.
但现在，当我跟更多决策者交流时，就会遇到另一类问题，而技术正好能帮上忙。
And now, obviously, when I talk to more decision makers, it's a it's a new set of problems that the technology can help solve.
就像你提到的，你们有潜在客户，但有时候就是没及时跟进，因为缺乏一套成熟的流程。
So as you mentioned, you have these leads, but sometimes you don't really reach out fast enough because you don't have that process.
就算有流程，人也可能在洗手间、生病了，或者不在岗。
Or even though you have that process, people are in the bathroom, they're sick, they're out.
对吧？
Right?
结果就没人去跟进后续了。
And then you don't have anyone following up.
我目前正在洛杉矶招一个外包人员，他们刚遇到另一个问题：他们的呼叫中心软件宕机了一个半小时。
I'm currently onboarding one contractor here in LA, and they just had another interesting issue where their call center software, had a buck for one and a half hours.
就是电话根本没转接出去，想象一下，客服团队现在连一个线索或一通电话都收不到。
It was just like not routing calls, and so imagine the call center team now not getting leads or any phone calls.
对吧？
Right?
这可是一个半小时的生产力和收入损失啊。你想过没有，比如AI能不能当个后备方案，让整个系统不仅更高效，还更抗得住突发状况？
Which is like an hour and a half of of lost productivity and revenue, and think about, you know, what are ways that, let's say, AI could be as a backup, right, to make your entire system not only more efficient, but also more resilient.
对吧？
Right?
所以，与其只靠增加人手，不如多配备一些团队成员——不管是真人还是AI数字员工，都能让整个系统越来越稳固、越来越高效。
So but just having more, I would say, team members, whether that's physical or or AI digital team members, you're just making the entire system more and more resilient and more efficient.
你们跟承包商合作，那他们对AI的反馈怎么样？
What has been the feedback that you because obviously you guys work with contractors.
AI当然有好有坏，对吧？
There's there's good and bad to AI, right?
我们一直想强调的是，有些电话确实还是得让真人来处理，这才是合适的用法。
So and one thing that we like to make sure we point out is, you know, there's there's good use cases to keep humans on for certain calls.
是的
Yep.
有些电话还是得靠人来处理。
People for certain calls.
也有一些情况其实没必要用人，可以重新分配，因为AI做得和人一样好，甚至更好。
And there's use cases that maybe shouldn't be or maybe they maybe they could re reallocate it because the AI can do as good or better than a a human can do.
你有没有什么见解？能给一些建议吗？
You have any sort of insights into what that looks like and what some recommendations would be?
有的。
Yes.
我们行业里最常见的担忧是：这些潜在客户资源非常昂贵。
So the the most common fear that we hear in the industry is, okay, we have these super expensive leads.
对吧？
Right?
我们不希望有什么东西的表现低于人类水平，那样会拉低我们的转化率。
We don't wanna have something that is not performing according to the human baseline, and so it's decreasing our set rate.
对吧？
Right?
这实际上让我们在亏钱。
And it's actually helping us lose money.
我对此的回应是，先找出当前业务中不够高效的部分。
The rebuttal I have for that is to identify areas of the business that are not optimal right now.
对吧？
Right?
比如，如果有人在接听每一个来电，像第二种情况那样，那很可能已经是最佳配置了，就不需要改动。
So if you have people that are answering every call like in the second, right, that's probably already an optimal setup, and then you shouldn't change it.
但有没有可能，比如更快地跟进潜在客户，然后安排预约呢？
However, are there opportunities, let's say, to call leads faster, right, And and try to schedule appointments.
如果数字客服或AI做不到，那还是可以由人工再打回去。
And if, let's say, the the digital call center agent or the AI is not able to do so, you can still call back as a human.
对吧？
Right?
你是不是星期天不上班？
Do you perhaps not work on Sundays.
却又想提供一种能让客户随时预约的服务。
But wanna offer a service that people can book appointments whenever they want to.
对吧？
Right?
那你可以先从让星期天变成工作日开始，现在星期天根本赚不到钱。
And then you just start by turning Sundays into a productive day, whereas now you're not making any business on Sundays.
对吧？
Right?
这些就是我建议的聪明做法：别光想着省钱，也要试试看怎么多赚钱，看看行不行。
And so these are ways that I would say you can be smart about not losing money, but taking a chance at making more money and seeing if it works.
如果有效，再慢慢扩展到其他方面。
And then if it works, you can expand into other things.
说到你提到的AI和人类的区别，我觉得AI特别适合那些流程稳定、可预测性高的场景。
Regarding your point about what the AI and the human is is like, where where they're different, I would say AI is really good where things are really predictable and where you have a clear, let's say, flow of things.
对吧？
Right?
我们最常见的做法是预约安排，这整个流程非常固定，需要从客户那里收集特定的信息来完成预约。
So the typical case that we do is appointment setting, and so this is a very predictable flow of questions and informations you have to gather from the customer to be able to set up the appointment.
整个对话大概三到五分钟。
It's like three, four, five minute conversation.
有趣的是，很多我们接触的大公司都遇到一个问题：客服人员并没有严格按流程执行。
And in a funny way, a lot of especially larger companies that we speak to, they have problems that the call center agents is not necessarily following these instructions.
对吧？
Right?
你可能会觉得，人类应该比AI强多了，毕竟他们眼前就摆着操作指南。
So you would think that, hey, a human should be like way better than the AI because they have this instruction in front of them.
但如果你在这行干了三四个月，就可能开始漏掉一些问题，觉得自己靠惯性就能搞定，不再按脚本来了。
But let's say you've been three, four months in the job, you might start to skip some questions because you think that you're doing stuff from muscle memory, right, and not according to the script.
所以某种程度上，有些人类的表现反而会低于基本标准。
And so in a way, that's where some humans even tap below the baseline.
对吧？
Right?
你以为自己已经达标了，其实并没有。
And and you think you have the baseline, but actually you don't.
所以这时候，你可以想想AI总是能严格遵守这个标准。
And so this is where you could think about AI always following that baseline.
对吧？
Right?
因为它就是个机器。
Because it's it's it's like a machine.
它会一丝不苟地按照脚本走。
Like, it it follows that that that script.
所以这些场景特别适合用AI。
And so those are really good use cases for the AI.
你提到有些人类会依赖肌肉记忆，或者默认这么做。
You mentioned that some some humans are going to use, like, muscle memory or Mhmm.
他们只是凭习惯去做，因为一来二去就成自然了。我们在企业里设计流程和规范时，经常遇到这种情况：创建流程的人一走，问题就来了。
Like, they're just going to do it because they like, it's so whenever we talk about building out processes and procedures in a business, a lot of times, the person that builds the process leaves a loss.
他们漏掉很多步骤，自己都没意识到，这些动作其实是靠肌肉记忆完成的。
They skip a lot of steps because they don't realize that they're doing it from muscle memory.
所以我们教大家的做法是，把流程写下来，或者口述出来，再找一个转录服务把内容整理成文字，然后交给别人去执行。
So what what we teach people to do is for them to write down the process or verbalize the process and have some transcription service, transcribe the process and then hand it off to someone else.
让别人来操作这个流程，看看哪里出了问题——因为你自己做惯了，忘了某些步骤，根本没意识到漏了。
Have them perform the process and figure out where you messed up because you you forgot to add a step because you're doing it by muscle memory.
你打算怎么解决这个问题？
How would you solve that problem?
我觉得唯一的方法是建立一个严谨的流程，确保每个人每次都严格遵守，同时还要设置第二道防线，也就是监督机制。
I think the only way to solve that problem is to have a really good process where you make sure that people always follow it, but then you have like a second line, which is essentially monitoring.
对吧？
Right?
比如把通话录音转成文字，然后检查每一个步骤是不是都被严格执行了。
So you transcribe the calls, and then you make sure that like these steps are always followed.
也许你可以用人工智能生成一个示例。
Maybe you can even use AI to just make an example.
把上个月的所有通话录音都上传到某个AI系统里，然后问它。
Just get all of last month's recording, you upload them into some AI, and you ask it.
对吧？
Right?
比如，告诉我有多少比例的通话遵循了这个流程，再给我一些没遵循的例子，顺便标出问题出在哪里。
Like, surface me what percentage of calls followed this process, and give me examples where it didn't, and maybe highlight okay.
听起来有点反乌托邦，但你可以特别标出那些最常不遵守流程的销售或团队，分析他们为什么不遵守，然后针对这些问题采取行动，跟他们沟通解决。
This sounds a little bit dystopian, but maybe highlight certain reps or groups of reps that most frequently did not follow this process, and like why they don't do it, so that you can then take that as an action item and address it with those people.
真巧，你提到这个，因为这几乎就是PhoneTAP的完整功能。
It's funny you mentioned that, because that's literally almost exactly what PhoneTAP does.
我正想说，你干脆直接买PhoneTAP就行了。
I was just gonna say, or you can just purchase PhoneTAP.
那再好不过了。
That'd be great.
对。
Yeah.
直接用PhoneTAP吧。
Just use PhoneTAP.
这正是我们作为企业主、经理或客服经理遇到的问题。
That's the problem that we were running into, like, as a business owner, or a manager or CSR manager.
比如你有15个客服，我不是说很多公司都这么多。
Like, if you have, you know, 15 CSRs, I'm not saying that a lot of companies do.
就算只有三个客服，你也不可能每天手动去听每通电话。
If you have three CSRs, you can't go back and listen to every single one of their calls all day long manually.
而且大多数人根本不想干这活。
Not only do you most people don't wanna do it.
我自己就特别受不了听电话录音。
I personally cannot stand listening to phone calls.
但这是很多客户对你们公司第一印象的来源。
But it's the first first impression that a lot of people have of your business.
所以我们常常忽视、忽视对他们的指导，培训也跟不上，结果他们没按流程来，我们反而还生气。
And so it's a very important role that we oftentimes, under under inspect, under teach, and then all of a sudden, we, like, do we train them properly half the time, and then we get mad at them because they're not following the process.
但往往只有出问题了我们才会发现。
But we only catch it
这个流程其实已经非常完善了。
You know, the process is, like, really fully dialed in.
是的。
Yeah.
没错。
Exactly.
说起来有点儿
It's kind
你想想看，这么安排其实挺讽刺的。
of ironic when you think about how you lay that out.
就像，确实如此。
It's like, yeah.
我们根本没把这些人的成功条件准备好，却还不明白为什么业务增长不起来。
We're not really doing a good job setting these people up for success, and then we don't we under we don't understand why our business isn't growing more.
那他们后来怎么办呢？
And then what what do they do?
他们只好去找营销公司，想办法多拉点客户。
They turn to marketing companies signing more leads.
这可是我最受不了的。
That's my favorite.
是的。
Yeah.
我能补充一点吗？
Can I add something on that?
没错。
So Yeah.
当然。
Of course.
首先，如果你还没用PhoneTAP，那真该用了。
First of all, of course, if you're not using PhoneTAP, you should.
杰布，凯文。
Job, Kevin.
回退到
Retracts in
这个问题到目前为止都没解决，原因就在于大数据。
The the reason the reason that this has so far been unsolved problem is because of big data.
对吧？
Right?
你每天产生这么多电话通话，这其实可以推广到你业务中的任何其他活动。
Like you're generating all these phone calls, and this could be abstracted to any other activity in your business.
对吧？
Right?
你每天产生这么多数据和活动，但如果不花掉宝贵的两小时，一头扎进细节里手动整理数据，根本没法分析。
You're generating all this data, you're generating all this activity, and there's no way for you to analyze this without spending two hours of your precious time every day to go into the weeds and pull together some data.
但这个问题实际上已经被我们的大型语言模型解决了。
But this is literally solved by our large language models.
对吧？
Right?
举个例子，作为企业主，我需要追踪我们的软件帮每个客户安排了多少预约。
So what we have done as an example, so in my business, I, as the business owner, have to track how many appointments our software is helping to schedule for each of our clients.
对吧？
Right?
以前这简直是个无解的难题。
Previously, this was like also a pretty intractable problem.
我得一个个登录系统，手动数有多少预约，然后试着理清这些数据。
I have to go into the deployments, like manually count how many appointments happened, and then try to make sense of that.
对吧？
Right?
现在我通过MCP把数据库和Entropic的Claude连起来了，设置了一个定时任务，每周日我陪老婆的时候，Claude就会自动分析所有客户的记录——包括通话转录之类的，统计出每周为每个客户安排了多少预约，然后按客户分周汇总好。
What I now have is I connected our database with Claude from Entropic, right, via an MCP, and what I did is to set up like a scheduled run where every Sunday, right, when I'm spending time with my wife, I have Claude go through all the different metrics for all customers, let's say all the transcripts, all the calls, right, and then have it analyze and pull together how many appointments we scheduled for them, and put them on a per customer per week basis.
所以我能看过去两周的趋势数据，每周日这些报告都会自动发到我的邮箱里。
So I see like the trending the the the the trending numbers from the last two weeks, and this is like a report that I can get every single Sunday in my email inbox.
而且在新的一周开始前，只要发现数据有问题，我就能直接通知相关同事去处理。
And then even before the week starts, I can like tag the relevant people to fix certain things if I notice things are wrong.
假设你还没用 PhoneTAP，
And so let's say you're not using PhoneTAP.
对吧？
Right?
你其实可以把呼叫中心的系统连上 Claude 或其他大模型，每隔几天自动分析数据，生成一份带洞察的报告发给你。
What you could do is you could actually connect your, let's say, call center software to something like Claude or another LLM, and have it perform this analysis on a structured way every couple of days and send you a report with the insights.
最棒的是，跟以前不一样，你不再需要一整个工程师团队来搞定这件事。
And the best thing about it is, unlike previously, you don't need a team of software engineers to execute this.
你只需要用一段文字描述清楚需求，就能实现我们说的‘氛围编码’。
Instead, you could have essentially just like a textual description of this, and and have And essentially just just what we call vibe code.
对吧？
Right?
就是这样。
And and just Yeah.
没错。
Exactly.
你只要用英文写下你的需求，AI 就会帮你搞定。
Just just type in English what you need, and and the AI will just do it.
而且还有很多其他应用场景，毕竟电话分析只是其中之一。
Well, and and there's so many use cases that too, because obviously that calls is one.
举个例子，我和Church以前参加同一个教练项目，那时候我还在做技术工作，他们教我们怎么设计评分表，把所有指标都列清楚。
But think about it from you know, Church and I were brought up on a we were in the same coaching program when I was still in the trades, and they taught us about scorecards and laying them all out and all stuff.
以前你得手动收集这些数据，或者只能依赖CRM里的数字——但我们都知道，深入看过CRM数据后，发现大部分都不准确。
Of course, you had to manually get all these numbers or you had to trust the numbers inside of your CRM, which we we know from just deep diving into CRM data, most of it's not accurate.
就算那些最用心、人手最充足的公司，数据也未必完全可靠，或者他们根本搞不懂这些数字在不同部门之间是怎么联动的。
Even the even the companies that try the hardest and have the most people trying to man it it's just not fully accurate, or they don't understand what how these numbers play across different, departments and stuff.
但现在，你可以轻松建立这些评分体系，清楚团队的表现，谁的转化率最高，每周分析利润表、材料成本和发票，各种数据都能一目了然。
But being able to build those scorecards and know how your team is doing, who who's who's converting the best, who's had you know, looking at that information, analyzing your P and L on a weekly basis, analyzing material costs and invoices, and just things.
在装修行业，最常见的盗窃方式之一就是有人拿几件零件，再悄悄多拿一两件。
Because one of the easiest ways of theft in the trades is someone goes and gets a couple parts, and maybe there's one or two more added on.
多拿的那几件记在公司账上，却放进自己的车里，拿去干私活。
That one or two goes on the company account and gets put in their personal vehicle for side jobs.
像这种事，你只要留心就能发现。
Just things like that that you can pay attention.
比如说，你就可以问：
Be like, hey.
我们明明只需要三件补货，为什么每次都买五件？
Why do we buy five of these when we only needed three to restock?
这类问题，你其实都能查得清楚。
And you can you can get these things.
是的。
Yeah.
这样你就能把管理做得更精细，不用靠猜，也不用莫名其妙地怀疑谁、指责谁，或者心里觉得有人偷东西，却只能憋着说些尴尬的话。
You can run you can run a tighter ship without having to be guessing or, just maybe, ghostly blaming someone, or it's like you don't really know, but you think they're stealing and having those weird conversations.
你其实可以真正知道，并且有实实在在的证据。
You can actually know and have some some proof.
这当然不是唯一能用的场景，但你们听的时候，应该会想到：哦，这个我能用上。
And it's not the, you know, it's not the only use case, but there's as you guys are listening to this and hearing, like, oh, I can connect this.
因为我和特什之前没怎么做过这方面的内容，很多人根本就觉得这太难了——我们聊Zapier都好几年了，大多数人一听就退避三舍。
Because we Tersh and I really haven't gone and done much content around this because a lot of people don't even like, we talked about Zapier for years, and most people was like, it's way over their head.
我们就想，行吧。
And we're like, okay.
那咱们还是从最基础的东西开始讲起。
Let's back up to, like, the basic level stuff.
现在你再把MCP服务器这些加进来，
And now you throw MCP servers and all these.
可还是有人连工作流是什么、怎么搭建都不懂，因为他们根本没像特什之前说的那样，建立系统化的流程。
I don't like, there's still like, don't understand workflows and how to make workflows happen because they don't have systemized processes like Tersh mentioned earlier.
所以一开始就把这些流程写下来、标准化，特别关键，因为只有这样，才能为自动化打下基础——无论是自动化结果、输出，还是你想分析的任何东西。真正让人惊讶的是，现在能分析、能自动处理的东西太多了，速度快得惊人，搞不好我们很快就要花更多钱在AI令牌和工具上了。
And that's why it's so crucial in the beginning to have a systemized written down process of how these things should get done because then that becomes the gateway to build things that can automate the outcomes or the outputs or the, you know, whatever you wanna analyze, but it is absolutely wild what can be analyzed and be done at a very high level very quickly, to the point where we're probably all gonna be spending more money on AI tokens and tools soon.
展开剩余字幕（还有 349 条）
到2030年左右，我们在这上面的花费可能会超过工资支出，因为它能快速分析大量数据。
Probably by 2030, we'll be spending more money on that than we will on wages because it's it it does analyze things, and you can get so much data so quickly.
但这也可能让人不知所措，因为回复动不动就是一大本书，你还得逐字读完。
But it also can be very overwhelming because the responses are all a book, and then you have to read through it.
有时候效果没那么理想，但这里面有太多机会能帮你更深入地了解自己的业务。
And, sometimes that doesn't work out as well, but it's there's so many opportunities to better understand your business.
所以正在听的朋友们，不妨去刷刷YouTube视频。
And so people listening to this, like, start watching YouTube videos.
什么是MCP服务器？该怎么操作这些工具？听到这些概念时就去了解，因为价值实在太大了。
What's MCP server, how to how to do these sorts of things as you hear them because the value is so immense.
你一定会学到很多东西。
And you're gonna learn a lot.
你肯定会感到困惑，但就像凯文说的，你以为开公司非得去上学不可？其实最好的学习方式就是直接上手干。
You're gonna be confused by a lot, but the only way to just like, Kevin, you've mentioned getting into a business and thought you had to go to school, The best way to learn is to jump in.
你好，家政服务行业的朋友们。
Hello, home service professionals.
你想把营销预算用到极致吗？
Do you want to maximize your marketing budget?
那你得用CallRail的来电追踪功能。
Then you need to be using call tracking from CallRail.
假设这个月你发现新线索突然增多了。
Let's say you've you've noticed a spike in new leads this month.
想知道是什么带来的，只要登录你的CallRail仪表盘就行。
To find out what's driving them, you just log in to your CallRail dashboard.
是我的新Facebook广告表现更好，还是付费搜索彻底碾压了其他渠道？又或者，是时候停掉报纸广告了？
Is my new Facebook ad outperforming my other ads, or is paid search knocking it out of the park, or maybe it's time to retire those newspaper ads.
因为CallRail会给每个营销渠道分配独立的电话号码，所以我清楚哪些渠道带来最好的线索，哪些该放弃。
Since CallRail assigns a unique phone number to each of my marketing efforts, I know where my best leads come from and which efforts to scrap.
CallRail会告诉你哪些营销渠道最划算，让你专心做业务。
CallRail shows you your most cost effective marketing channels so you can get back to business.
去callrail.com/sbmpod免费试用吧。
Go to callrail.com/sbmpod and try it for free.
有个问题想问一下。
Quick question.
上个月你漏接了多少通电话？
How many calls did you miss last month?
不是接了多少，是漏接了多少？
Not how many you answered, how many did you miss?
在家居服务行业，大约有三成的来电没人接。
Across the home services industry, about 30% of inbound calls go unanswered.
这些可不只是随便问问的人。
And those aren't just tire kickers.
这些都是随时准备找第一家能预约的商家下单的客户。
Those are customers ready to give their business to whoever books first.
Breezy 用AI客服接听每通电话，帮你预约服务，还能立刻跟进。
Breezy fixes that with AI agents that answer every call, they book the jobs for you, and they follow-up instantly.
别再一睁眼看到的是漏接来电，而要一睁眼看到的是已预约的单子。
So instead of waking up to missed calls, wake up to book jobs.
如果你想看看自己现在到底丢了多少单，去 getbreezyapp.com，用代码 sbm，他们会送你 500 美元用于体验 Breezy。
If you wanna see how many jobs you're losing right now, go to getbreezyapp.com, use code s b m, and they'll give you $500 towards Breezy on us.
好了，我们继续回到节目。
Now back to the episode.
我想补充一点。
I I would I would like to add.
首先，拥有流程、愿景和想法真的特别重要。
So first of all, having a process and a vision and an idea is absolutely important.
我敢说你们九成九都知道，有本特别棒的书叫《EOS：企业运营系统》，讲的是怎么经营公司，推荐大家读一读。
I I I'm 99% sure you guys know this, but there's this great book called EOS, Entrepreneur Operating System, of how to run your company, so people should read that.
但它讲的核心是，作为老板，你怎么从日常事务里抽身出来？
But what it talks about is how do you detach yourself as the business owner from the business, right?
不是事事亲力亲为，因为那样根本没法扩张，关键是把合适的人放在合适的位置上，然后放手让他们干。
Where it's not you doing all the work, because that's how you don't scale, but essentially how you put the right people in the right place, and then have them cook.
而你作为老板，主要负责分配资源，同时也要做最终的把关，而AI完全可以帮你承担大部分的监控工作。
But then you as a business owner, you are the capital allocator, and of course, you're the ultimate monitoring instance, and AI can do a lot of the monitoring.
CRM之类的系统不准确，就是因为得手动录入数据。
The reason why CRMs or something are not accurate is because you have to do data entry.
对吧？
Right?
比如说打完一个电话，得把信息记下来，但我可能就忘了。
So let's say a call happens, have to end the data somewhere, and maybe I forget to do it.
但有了AI，这还是第一次，能直接从源头获取数据。
But with AI, it's the first time where you can have something go straight to the source.
你可以直接连上你的财务软件、银行软件、通话系统、会议录音，还有销售工具，让数据直接流入大数据池，中间完全不需要任何中介软件。
You can literally connect it to your accounting software, to your banking software, right, to your call software, to let's say your meeting recorders, to your sales software, right, and then have it directly sit to that big data without like any software in between.
我还要挑战一下，说你非得先搞懂什么是MCP。
And I would also challenge the fact that you have to inform yourself what an MCP is.
其实，根本不用管什么MCP这个术语。
In fact, forget about, like, the term MCP.
直接去用Claude Co-Work，或者那个聊天机器人、Claude Code就行。
Just go to Claude co work or the chatbot or Claude code.
别管什么代码了，也别被那些焦虑困扰，直接打开就行。
Forget the the word code that and and and and the anxieties that caused, just just open it.
这就像是个文本框，你只要用英语打字就行。
It's like a text box where you can like type English.
对吧？
Right?
忘了代码这个词，直接用英语告诉它你需要什么。
Forget the word code, and just let it know what you need in English.
再说一遍，别想代码的事，AI会告诉你该怎么做，才能实现你想要的业务成果。
Again, forgetting the word code, and then it will tell you, like the AI will guide you what you need to do to be able to get certain business outcomes.
如果有人觉得这背后其实有代码，或者觉得太难了，给我发封邮件。
And if anybody thinks, oh, there was actually code involved, or it's too hard, send me an email.
但如果你没这么想，我向你保证，这真的只是用英语描述你想要的业务目标和结果，AI会一步步引导你、帮你完成所有操作。
But but if not, like, I I promise you that it's actually just using English to describe business outcomes and and results that you need, and the AI will guide you and perform all the steps so that you can get set up.
它会告诉你需要做什么，然后你只要给它授权，让它能访问相应的系统就行了。
And it will let you know what you have to do, and then you will just give it the right permissions to give it access to the right systems.
我觉得人们在使用Claude Code和Cowork时最大的误解，就是觉得不会编程就用不了。
I think that's the biggest thing that people get confused with Claude Code and and Cowork is they assume that if you don't know how to code, then you can't use it.
其实完全不是这样。
And that's not that's not the case.
你只需要直接问它：嘿。
Literally, you just ask it, hey.
帮我提20个关于x y z这个主题的问题，我需要你帮忙。
Ask me 20 questions that pertain to x y and z topic that I need I need you to help me with.
然后它就会生成一个完整的资料库，真的。
And so then it's it creates a, I mean, a full repository of information that Yeah.
完全基于你给出的回答。
Based off the answers that you gave it.
我觉得这还是很多人常见的误解。
I I think that that's the other common misconception that people have.
不过你说得一点都没错。
And and you're exactly right, though.
直接开始就行。
Just get started.
先问你的问题，让它也来问你，然后慢慢推进。
Just start asking your questions and and let it ask you questions and, kinda go from there.
我觉得我们大多数人都是这么开始的。
That I think that's how most of us got started with it.
我经常用它来处理内容，因为它能想出点子、列提纲什么的，但我从不直接用，这对我来说是个个人习惯。
I do that with content all the time, like because it'll it'll it can come up with ideas and outlines and stuff, but I I don't ever It's it's a personal thing.
有些人可能觉得随便丢点AI生成的内容就挺好。
Some people will probably be fine just throwing some AI stuff out.
我喜欢先给它一些背景信息。
I like giving it context.
我更喜欢它先问我问题。
I like it asking me questions first.
在发布任何东西之前，我总希望它能先贴近我的表达方式、我的想法。
I always want it to be in my voice or my thoughts or whatever before I post something.
AI能帮我写出来、排好格式，让内容从头到尾连贯顺畅，而不是东一榔头西一棒子。
AI can help me write it and format it and put it in a way that it gets from a to z so it actually, like, makes sense instead of, like, bouncing all over the place.
但最终要做出真正有成效的东西，还得靠我们这些创业者、企业领袖和内容创作者的经验。
But have it it it needs that experience that we have as entrepreneurs, business leaders, content creators to to create something that's really actually gonna produce well.
有时候人们觉得，哦，AI能帮我写篇博客。
And I think sometimes where it's like, oh, this can create a this can create a I'll use blog for me.
不错。
Cool.
博客文章。
Blog post.
搞定。
Done.
好像只是完成了一个任务清单，但问题是，写出来的内容太泛泛了。
Like, it's just like a it's a checkbox item, but the problem is the blog is really generic.
缺乏真正的见解，也没有人情味。
It doesn't have any real insights, real human connection pieces.
它缺少一个连贯的故事把内容串起来。
Doesn't doesn't have a story to bridge it together.
我觉得这往往是被忽略的关键点。
And I think that's sometimes the missing pieces.
我们一看，哦，能多做点，就不停地赶工。
Like, we look at is, oh, we can do more, so let's just keep boom boom.
尤其是当人们开始频繁使用这些工具时，更容易陷入这种状态。
Let's just keep getting stuff out the door, especially when people start becoming power users of this stuff.
但这样一来，内容反而失去了原有的价值，也少了真正能吸引人的东西。
But it loses it starts losing, I think, some of the value and some of the stuff that people actually want to pay attention to.
我们只是因为现在能做，就去做，却没想过：这真的解决了我们业务中的问题吗？
And we just like because we can do it now, we do it instead of is it solving a is it solving a problem in our business?
如果没解决，那我们为什么还要花时间在这上面？
And if not, why are we focused on it?
我们只是因为觉得‘这挺酷’，就盲目投入，浪费了时间。
Why are we just wasting time on something that can oh, it's cool.
是的
Yeah.
有很多事情
There's a lot
看起来挺酷
of things that are
但它能带来收入吗？
cool, but I mean, is it producing revenue?
能降低成本吗？
Is it saving on the bottom line?
能推动你的业务前进吗？
Is it doing something that's gonna move your business forward?
还是你只是在浪费时间，做些自以为很酷的东西。
Or you're just wasting your time building some cool shit that you think is cool.
有时候
I mean, sometimes
做这件事也很有意思。
it's cool doing that too.
我是说，凯文，你可能不知道。
Well, I mean So, Kevin, you don't know this.
特什用Zapier搭建了这个系统，我没法访问，但他每天都会给我发一个爸爸笑话。
Tersh built this through Zapier, and he I don't have access to it, but I get a daily dad joke every day.
他还把笑话发到Facebook上，就在今天早上，我们刚遇到一件事。
And he he also posts it to Facebook, and we literally had someone earlier this morning.
我们正在谈合作，对方说，看到这个笑话出现在动态里，真的能让人心情变好。
We were having a partnership conversation, and, it brightens he said it brightens my day to see that in my feed.
所以，你也可以自动化一些特别傻的事。
So you can automate some really dumb shit too.
是的。
Yeah.
但这么做确实挺有意思的，你知道的。
But it can be fun, you know.
所以这其中是有来有往的，有些东西拿来自己玩玩，其实也挺有意思的。
So there's there's a give and take it, and some of that stuff is is cool for just, you know, personal playtime.
但说到商业应用，你提到的那些——分析数据、整合数据源之类的，你得问问自己：我到底想解决什么问题？
But, if as far as business application, what you're talking about, like, analyzing and connecting your data sources and stuff, what is the ask yourself, what's the problem that I'm trying to solve?
这样你才能不跑偏，因为这些花哨的工具太多了，很容易让人分心，用法简直无穷无尽。
So you can actually stay on topic because you can get distracted very easily with all these fancy tools because it's endless use cases.
凯文，你觉得语音AI和文字AI在不久的将来会怎么发展？
Kevin, where do you see AI voice and AI text messaging in the near future?
对。
Yes.
我觉得未来，它会成为一个非常重要的团队成员。
So I think the way I see it is that in in the future, this will be very important team member.
没错。
Right?
对大多数中型甚至大型的承包类企业来说尤其如此，因为它能更有效地解决一些棘手的问题。
For for most, like, mid sized larger contracting businesses, because it just helps to solve some of these pain points in a in a better way.
我们现在的产品是这样的，有这些AI语音助手，能接电话，一次能同时处理一百通电话，还能在潜在客户一进来后的七秒内主动拨出去。
The way we currently have our product is that we have these AI voice agents that can, you know, take calls, right, can do a 100 parallel calls, can also dial leads within seven seconds of them coming in.
如果打语音电话没人接，我们就发条短信，对方可以用聊天的方式直接预约时间。
And then if we have, you know, a a voice call and somebody doesn't pick up, we send them like a text message, and then they can, in a conversational way, schedule an appointment.
目前我们还没有但正在开发的是自动化营销活动功能。
What we currently don't have that we're working towards is something where you can create automated campaigns.
比如说，针对一个潜在客户，你可以设置：第一天打八次电话，发一条短信。
So let's say, given a lead, you can say, hey, call them, let's say, I don't know, eight times on day one, send one text message.
第二天打三次电话，再发一条短信。
On day two, call them three times, send one text message.
然后就这样持续下去，持续三十天左右。
And then so forth, you know, for for the next thirty days.
因为有时候你根本没法一次就约到客户，特别是某些行业。
Because sometimes you don't really book an appointment with someone, you know, especially for some businesses.
对吧？
Right?
比如屋顶维修这种行业，可能得持续到第45天才约得到人。
Like like, I don't know, roofing, until day 45.
所以你真正需要的是能持续跟进客户的系统，而这一点其实人类也能做。
So you actually want to have something that can continuously follow-up with customers, and so this is something that let's say humans can also do.
打第十次电话确实挺烦的，但AI根本不在乎。
It's like pretty annoying, right, to call someone for the tenth time, but AI doesn't really care.
它只管不停地拨号，一旦接通了，就直接安排预约，或者转接给真人团队。
It just places to call, and then if you reach someone, it actually schedules schedules an appointment or creates a warm transfer for the human team.
别把AI想成完全取代人的东西，它更适合干那些枯燥乏味、但又必须有人做的活儿。
So think about AI not as something that totally replaces people, but as something that can do really boring stuff, but that you should or want to have done.
从长远来看，六个月后，人类员工反而会感激有这样的系统，因为他们再也不用干这些烦人的事了。
And and in a way, it should be something that after six months, like, the the human qualities are happy to have on their end, because they don't have to do like annoying stuff anymore.
这些事，恰恰是AI最擅长做的。
And it's something that that AI can do.
你觉得用AI语音做外呼去搞骚扰、搞恶心人的事，会不会导致AI语音被限制？
Do you think that people using AI voice for outbounding for nefarious reasons, annoying reasons, is going to put limitations on AI voice?
是的。
Yes.
当然。
For sure.
对吧？
Right?
现在有些人会用AI打恶作剧电话。
So you have, you know, people doing prank calls.
AI的优势在于，一天能打十万通电话，专门用来跟进老客户或者重复营销。
The the advantage of AI is that you can, you know, do a 100,000 dials per day, you know, for for rehash campaigns or to existing leads.
但这也导致全国范围内的电话数量暴增，所以我们才需要DNC这样的屏蔽名单。
But of course, it also leads to a fact that now you have just way more calls being placed in the country, which is why we have something like a DNC database.
对吧？
Right?
如果有人告诉我们一次别再打了，我们就会把他们加入内部屏蔽名单。
If someone tells us one time that, like, do not call me again, then we put them in our internal database.
这不是额外的东西。
It's not something extra.
我们内部早就有了这个机制。
Like, we we have that internally.
下次我们的客户用我们的系统拨打他们的电话时，系统会直接拦截。
And then the next time one of our clients tries to call them using our technology, we'll actually block that call.
对吧？
Right?
真厉害。
So Wow.
我们能阻止一个人在三十天内被拨打三十次吗？
Can we prevent someone being called 30 times over thirty days?
不能。
No.
这更像是一个功能，而不是限制，因为人类也会这么做。
It's more of a feature than a buck because humans do that too.
但如果有人明确说一次，别再给我打电话了，我们确实会尊重这个要求，以后再也不打给他们。
But if someone says like one time, like, do not call me again, we will actually like respect that and not call them again.
你是不是也跟我一样，正面临同样的困扰？我知道乔什也是，每天接到十几通、甚至三十通AI电话，都是关于贷款审批、信用审核这类事？
So are you facing the same dilemma that I'm facing, and I know Josh is too, where you get, I don't know, fifteen, thirty AI phone calls a day for, what is it, financing, underwriting?
贷款相关的事。
Loan thing.
是的。
Yeah.
信用审核。
Underwriting.
是的。
Yeah.
这个
This
你已经连续清醒35个小时了
is this is You've been up for 35
部门。
department.
我都在想，这是哪家公司的核保？
I'm like, underwriting of who?
你把公司名字写上去。
Put your company name in there.
对。
Yeah.
每次都是这一套。
It's it's the same thing every time.
我接到这种电话就想，我是不是
I get this, I'm like, did I
不小心注册了什么贷款服务？
did I accidentally sign up for some loan thing or something?
我怎么会被加到这个名单里的？
Like, how do I even get this on this list?
我的谷歌资料肯定出问题了。
That my Google profile is definitely screwed up.
这倒是个好事。
Good thing.
这就涉及到人工智能的合理使用了。
That's where, like, the responsible use of AI comes in.
我们合作的客户通常是那些通过自己网站、Facebook广告或谷歌广告主动获取了线索的人，客户已经表现出明确兴趣，正等着有人回电话，就像乔什开头说的那样。
So the the businesses we like to work with are those that just received a lead on their own website, for example, or on like a Facebook ad or a Google ad, and then those customers are actively interested and and waiting on a call, like like Josh described in the beginning.
在这种情况下，人工智能并不是在推销，只是帮忙预约时间。
And so those are cases where AI is not really selling anything, but it's like scheduling an appointment.
对吧？
Right?
我们也不想太强推。
And then we're like not trying to be too pushy.
那些只想疯狂发广告的公司，我们一般也不会合作，比如客户根本没意识到自己对园艺、换窗户、屋顶维修或者厨房翻新感兴趣。
And businesses that, you know, just wanna spam customers are usually those that we also don't work with, let's say where leads don't even know that they're, like, interested, let's say, in in landscaping or window replacement or or roofing or kitchen refacing.
对吧？
Right?
这些情况下，不管是用AI还是用人，成功率都特别低。
And those are cases where also your success rate with AI as with humans is like super super low.
是的。
Yeah.
理想情况下，你应该负责任地使用AI，确保用户知情同意，只联系那些真正感兴趣的人，这样AI就能成为强大的助力，而不是滥用它去打扰那些根本不知道你这家企业存在的客户。
So ideally you use you like use AI in a responsible fashion, you make sure that people are consenting, you only contact those that are interested, and then AI is like a superpower, and you don't misuse it for for contacting people that don't know that you as a business even like
对。
Yeah.
甚至存在与否都不清楚的人。
Exist or or yeah.
你觉得做这些事的人，最终会不会给整个行业带来麻烦？
Do you think that that people that are doing this are going to ultimately cause issues with the rest of the industry?
我感觉确实会。
Like, I feel like Yeah.
你要是惹恼了对的人或者错的人，突然间他们就可能推动立法，禁止AI外呼之类的措施？
You annoy the right person or the wrong person, and then all of a sudden, they're putting up legislation to ban AI outbound calling or something like that?
是的。
Yeah.
希望不会到那一步。
Hopefully not.
对吧？
Right?
我的意思是，确实如此。
I mean Yeah.
这就像加密货币行业，有些人在别的方面也干着违法的事，比如挪用资金、把钱转到其他生意里，这种行为会拖累整个行业。
It it's like, the parallel you could draw is like the crypto industry where you have some guys that are, you know, even for other reasons, like embezzling money and stuff, or funneling money into other business, and this just puts a dent on the entire industry.
所以，大多数企业应该负责任地使用AI，而那些滥用的人，问题更多出在他们的商业行为上，而不是AI本身，希望如此。
And so it does, you know, it should be the case that most business are using it in a responsible way, and those that are using it in non responsible ways, it reflects less on AI hopefully, than on their business practices and
当然了。
Oh, for sure.
他们的态度。
Their attitudes.
我推测，这些人其他方面也可能不太靠谱。
And and I would hypothesize that those guys, they also are maybe a little bit fishy in other ways.
对吧？
Right?
他们干这种事， probably 不是第一次做这种不地道的事了。
Like, it's probably not AI is probably not the first time they're doing some fishy stuff.
没错。
Oh, yeah.
确实如此。
Yeah.
他们私下八成还在干些见不得人的勾当，但希望人们更多是怪他们做人有问题，而不是怪AI本身。
They're probably doing some some fishy stuff on the side, and hopefully, it reflects less on AI than, like, how they're using it.
当然。
Absolutely.
不对。
No.
我同意。
I agree.
如果你真把社保号给他们，那就是自找麻烦，毕竟他们连这种明目张胆的勾当都敢干，显然不是什么靠得住的人。
That, like, if you if you actually tried if you gave those people your Social Security number, you're asking for it because you know, like, if they're already doing this nefarious behavior upfront, like, you know they're not gonna be the most trustworthy people.
但生态圈里的人其实还是能做点什么的。
But but there's, like, something that people in the ecosystem can do.
比如我们的软件就特意设计成，AI每天联系同一个住户最多不超过三次。
So for example, what our software does, it actually by design prohibits from our AI contacting a homeowner more than three times a day.
我们这么做是有道理的，因为如果你用同一个号码一天内频繁联系某人超过三次，AT&T、Verizon和T-Mobile这些运营商很快就会把你的号码标记为垃圾电话。
So we do that for logical reasons, because if you contact someone, and this is like a general point, more than three times a day using the same phone number, AT and T, Verizon, and T Mobile will probably mark your number as spam within like a couple of moments.
对吧？
Right?
所以我们还自己开发了一套号码监控系统，其中一条重要经验就是：一天内别给同一个人打超过三次电话。
And so we something we do is we have also our own like number help monitoring software, and one of the lessons we learned is to not call people more than three times a day.
所以我们系统里本身就设置了这个限制。
So this is something that's built into our system.
对吧？
Right?
如果你想要同一个客户打第四次，系统会直接阻止这次通话。
If you want to contact like the same customer four times, we will actually like block the fourth call.
对吧？
Right?
我们就是想通过这种方式，防止AI被滥用。
And this is how we're trying to help the AI being misused.
当初会出这个问题，其实原因没那么严重——以前我们确实会打六次。
The original reason that this was even the problem is way more benign, and that in the in the past we did call people like six times.
为什么呢？
Why?
因为同一个潜在客户，同一个电话号码，在客户管理系统里重复出现了两三次。
Because the same lead, the same phone number was in the CRM like two or three times.
对吧？
Right?
你看到同一个电话号码在客户管理系统里出现了三次。
You had the same phone number being like three times in the CRM.
其实是同一个人，只是地址不同，或者名字拼写有点小差别。
It's like the same person, but like maybe with different address or like the first name spelled slightly incorrectly.
这更多是客户管理系统数据不干净的问题，而不是人工智能的问题。
And so this is more a problem of CRM data integrity than like AI.
但我们当然还是要从客户管理系统里提取这些数据。
But of course, we're trying to pull all this data from the CRM.
对吧？
Right?
结果同一个号码出现了三次，如果每个都打两次，那自然就打了六次。
And we got the same number like three times, and then if we call each person like two times, then of course you call them like six times.
但正是在这种情况下，像我们这样的人，或者其他团队，可以想办法绕开这个问题。
But this is where people like us, people like other people, we can build workarounds.
对吧？
Right?
比如屏蔽重复号码、去重之类的，肯定有办法解决。
Like blocking the same number, deduplicating and stuff, and and so there are definitely ways
数据和CRM系统的问题确实非常严重。
of data and CRMs would be a really, really is that in that in,
没错。
like Yeah.
我之前待过一家空调公司，一开始我们录入客户名字都是名在前、姓在后，后来因为QuickBooks系统的一些原因，改成了姓在前、名在后。
You know, there was a time the last HVAC company I worked for, we had because for one reason, we would always the way we started putting names in, was, like, first name, last name, and then we changed it because QuickBooks and some so some reason with QuickBooks and stuff, we did last can comma first name.
现在你要是想导出名单做邮件营销，根本没法用合并字段，数据全乱了。
Well, now you got, like, when you try to export a list to maybe go do an email marketing campaign or stuff, you got like, you can't use merge tags because it's a mess of data.
是啊。
Yeah.
姓写完了，要是夫妻俩名字，还得加个&符号，更乱了。
Last name, and then you got the ampersand if there's a two first names for the husband and wife.
我们经常为客户做这种事。
And it's just we do this all the time for clients of ours.
我们会导出名单，打电话提醒客户做保养和检查，好再上门服务。
We pull lists for, you know, outbounding to get tune ups and inspections and get back in their our our customers' customer's house.
我们不得不专门开发了一个自动化工具，处理各种小问题来清理这些数据，因为手动改实在太麻烦了。
And we had to literally build an entire agent that fix it with all of the little use cases to fix these lists because it sucks to fix.
很多人根本没意识到问题有多严重。
And people don't they don't even really not nuts.
你难道不能直接导出名单吗？
Can't you just pull listen.
发个短信就行了。
Send the text message.
是的。
Yeah.
你可以这么做，但到时候你会收到一堆像‘嘿，布莱塞特’或者‘嘿，对’这样的回复。
You could, but then you're gonna get like, hey, Blissett or hey Yeah.
嘿，吴，你应该叫你姓氏，而不是嘿，凯文。
Hey, Wu, you know, your last name instead of hey, Kevin.
这样一来，明显是自动化出了问题，根本拿不到回应。
And then it's like, well, it's clearly an automation that's not working well, and you're not gonna get a response rate.
你很快就会把整个名单都耗尽。
You're gonna you're gonna blow through your list.
所有人都会回复退订，到时候你还会被运营商拉黑。
They're all gonna say stop or unsubscribe, and then you're gonna hurt yourself with the carriers.
所以，我接下来要说的可能比较实际。
And the deal So maybe what I'm about to say now is grounded.
听着像硅谷那一套。
Sounds like Silicon Valley.
对吧？
Right?
可要是换作来自佐治亚，或者美国中部的人，他们可能根本听不懂这人在说什么。
And someone from, I don't know, like Georgia or in the middle of Georgia or where I'm sometimes like, what the fuck is this guy talking about?
对吧？
Right?
但我还是换个说法吧。
But let me just say it other ways.
我每次都说这个。
I say it every time.
我真不知道你其实明白。
I have no idea you're Yeah.
一半的人试着用AI来解决这个问题，看看行不行。
Half the Try try to solve this this problem with AI and see if it works.
对吧？
Right?
就花一个小时试试。
Like, try try for one hour.
去云平台，连上你的HubSpot或者任何CRM系统，找出里面所有电话号码相同的客户。
Like, go into cloud, connect it, let's say, your HubSpot or whatever CRM, and say, figure out every single customer that is in there with the same phone number.
只要看到同一个电话号码出现两次，就把所有相关记录都找出来。
Like, anytime you see the same phone number twice, give me all those instances.
第一步。
Step number one.
对吧？
Right?
看看这样行不行。
See if that works.
然后想办法把数据合并起来。
Then merge the data in some way.
对吧？
Right?
弄清楚为什么会这样。
Figure out why that's the case.
比如你可能有来自Facebook的线索，还有第三方供应商的，其实是同一个人被录入了两次，试着把它们合并，再让AI试着处理五条记录，看看能不能成功，最后再用AI清理整个CRM系统。
Maybe you have Facebook leads and from a third party vendor, right, it's like the same guy just got entered two times, and try to merge it, and then give AI maybe permission to perform this act for like five leads to see if it can work, and then actually clean your entire CRM with AI.
对有些人来说，这听起来可能像硅谷的疯狂想法，但不妨试试看。
And maybe it sounds like crazy Silicon Valley to to most or some people here, just give it a shot.
如果真能行，那可是相当大的价值。
And like, if it works, it could be Really, a good value.
对吧？
Right?
不管你在哪里做生意，这都能带来巨大好处，到时候别忘了谢我。
It's a huge benefit to your business wherever you are, and and and thank me later.
我觉得就这些了。
I think that's it.
我是个打工人。
I'm a slave.
管用。
Works.
是的。
Yeah.
我完全同意。
I agree, a 100%.
这些正是我们当老板的平时抱怨的事情。
And it's those things that, like, we as business owners complain about.
每次都要重做一遍，真是烦死了。
It's like, ugh, we have to do this again.
这活儿太枯燥了。
It's such a monotonous task.
只要一冒出这种念头，马上问自己：这事儿能用AI搞定吗？
Like, as soon as you think that thought, immediately say, can I use AI to do this task?
十次里有九次，我都会说：没错，应该能行。
And nine times out of 10, I'd I'd say, yeah, you probably can.
我得提醒你，Claude 无法通过 Chrome 扩展访问 Stripe。
I will tell you, Claude cannot access Stripe using Chrome, the Chrome extension.
我试过用它来整理和解决一些 Stripe 的问题，结果它说：抱歉，办不到。
It's that's one thing that like, I've tried having it, like, organized and fixed some Stripe issues, and it's like, sorry.
我们没法访问那个。
We can't access that.
好的。
Okay.
我又要用那三个禁用词了，MCP。
I'm gonna I'm gonna use the three forbidden words again, m c P.
不过，别管我刚才说的，你直接去CrowdCode，试试连接Stripe，看能不能行。
But again, like, forget that I said that, but just go and crowd code, say connect it to Stripe, and see if it works.
我这边已经连好了Stripe，可以做自动化的客户仪表盘。
Like, I I have Stripe connected, so I can build automated, like, customer says dashboards.
比如，提醒客户付款逾期了。
Like, hey, this customer is late on a payment.
对吧？
Right?
哦。
Oh.
甚至可以通过云端给这位客户发封邮件，提醒他们付款迟了。
Maybe even send that customer an email via cloud, you know, like, hey, you're late.
设置成定时任务，每两天自动运行一次，确保及时收款，别让客户拖欠账单。
Do it as a scheduled run, you know, every two days to make sure you're getting paid and and not have customers that are late on invoices.
这真是个好主意。
That's a great that's a great idea.
毕竟，就算不用Stripe，如果你用的是ServiceTitan、Housecall Pro、Breezy这类现场服务管理或客户管理系统，也一样能跟踪那些拖欠款项的客户——账款拖太久，生意根本撑不住。
Because, I mean, even, like, using it not necessarily from Stripe, but, like, if you have ServiceTitan or or Housecall Pro or Breezy or any of these these FSMs and or CRMs, and you wanna stay on top of people who who I mean, you can't have an extended accounts receivable account, like, and stay in business Yeah.
拖得太久可不行。
For too long.
所以，别想太多，先花三十分钟试试看。
So And just just just try it for, you know, thirty minutes.
去Claude那里下载试试。
Go and Claude, like download it.
如果不行，咱们再回头做平常周日该做的事，陪陪孩子就好。
If it doesn't work, you know you know, let let's let's go back to doing the normal Sunday activities, know, spending time with kids.
就试三十分钟吧。
Just try for thirty minutes.
最坏的情况就是没效果，那也没关系，大不了再陪孩子呗。
The the worst case scenario doesn't work, like, fair enough, spending time with my kids again.
最好的情况是真有效，那你就能早点收到钱。
The best case is it it does work, and then you're getting paid earlier.
对吧？
Right?
银行里钱多了，你就停不下来了，还想继续自动化更多流程，越搞越上瘾，生意也越做越高效。
And you have more money in your bank account, and then, like, you you can't stop, then you wanna automate more things and get Then you get addicted to business even more efficient.
是的。
Yeah.
没错。
Yeah.
我觉得这正是当初这个播客开始的原因。
I think that's how the podcast originally got started.
我们就是越来越沉迷于把能自动化的事情都自动化，然后就能分享这些经验了。
Like, we just we're just getting addicted to automating things as much as possible, and then we can start sharing about it.
没错。
But, yeah.
不是的。
No.
我完全同意。
I I agree a 100%.
凯文，我非常珍惜我们刚才聊的这些内容。
Kevin, I really value everything that we've talked about here.
我确实想问你一下，乔什，你可能记得答案，但那种不属于你的奶酪，你叫它什么？
I I do wanna ask you, Josh, you may remember the answer to this, but what's what do you call what do you call a cheese that's not yours?
我知道答案。
I know the answer.
你觉得呢，凯文？
What do you think it is, Kevin?
我不知道。
I don't know.
我得靠乔什帮帮我。
I need I need Josh to help me out.
那是‘不是你的奶酪’（Nacho Cheese）。
It's nacho cheese.
这笑话挺烂的。
That was the bad
今天早上那个笑话。
joke this morning.
这笑话不错。
I It's a good one.
今天是录播客的日子。
It was a a podcast recording day.
我特意认真看了这一条。
I made sure I, like, fully read this one.
我好像记得她曾经问过我。
I'm like, thought she asked me at some point.
是的。
Yeah.
对。
Yeah.
没错。
Yeah.
这笑话不错。
It's a good one.
我今晚去酒吧就用这个。
I'm gonna use it at the bar tonight.
我之前说自动化的时候可不是在开玩笑。
I wasn't kidding when I study automated it.
真的每天都在用，连周末都不例外，它总是
Like, literally every single day, even the weekends, like, it's
是的
Yeah.
对
Yeah.
你明白的。
You get them.
对
Yeah.
凯文，谢谢你陪我们。
Kevin, we appreciate you hanging out with us.
乔什，你还有什么想问的吗？
Josh, is there anything else you wanna ask?
凯文，你还有什么要补充的吗？
Kevin, is there anything you wanna add?
都没问题。
All good.
特别有意思。
It's really fun.
凯文，大家最方便通过哪里联系你呢？
What's the best place for people to reach out to you, Kevin?
在我们网站上。
It's on our website.
对吧？
Right?
他们可以主动联系我们。
They can reach out.
我们有个表单可以提交。
We have a form submission.
他们可以试试AI功能，获取更多信息。
They can play around with AI, get some more information.
另外，就在我们的网站 leapingai.com 上。
And, yeah, just on our website, leapingai.com.
天哪。
Oh, man.
我们非常感谢。
We appreciate it.
如果任何人有任何问题，随时联系凯文都没问题。
And if anybody has any questions at all, don't hesitate to reach out to Kevin.
好了，祝大家这一周过得愉快又平安，我们下次再聊。
And with that being said, hope you have a wonderful and safe week until we talk to you again next time.
我们回头见。
We'll see you later.
再见。
Bye.
感谢收听《服务业务精通》这一期节目。
Thank you for listening to this episode of Service Business Mastery.
现在，你已经掌握了这场精彩对话中提供的关键商业建议，离成为你梦想中的成功老板又近了一步。
Now that you are equipped with essential business advice from this impactful conversation, you are one step closer to becoming the successful owner of your dreams.
如果这期节目对你的事业有帮助，别忘了订阅节目，留下评分，也分享给其他老板。
If this episode has been helpful to your business journey, don't forget to subscribe to the show, leave a rating, and share it with other owners as well.
更多内容请访问 servicebusinessmastery.com。
Visit servicebusinessmastery.com to learn more.
营销确实不容易。
Marketing is hard.
但我告诉你一个小秘诀。
But I'll tell you a little secret.
其实没那么难。
It doesn't have to be.
我来跟你讲个重点。
Let me point something out.
你现在正在听播客，对吧？你很喜欢主持人，会主动找来下载，开车、健身、做饭，甚至上厕所的时候都在听。
You're listening to a podcast right now and it's great, you love the host, you seek it out and download it, you listen to it while driving, working out, cooking, even going to the bathroom.
播客就像是个贴心的伙伴。
Podcasts are a pretty close companion.
这是一则播客广告。
And this is a podcast ad.
我吸引到你了吗？
Did I get your attention?
通过 LibsynAds 的播客广告，你可以触达像你这样的优质听众。
You can reach great listeners like yourself with podcast advertising from LibsynAds.
你可以从数百个热门播客中选择主播推荐的形式，也可以像这样投放预制作的广告，覆盖数千档节目，精准抵达目标听众在他们最爱的播客里。
Choose from hundreds of top podcasts offering host endorsements or run a pre produced ad like this one across thousands of shows to reach your target audience in their favorite podcasts with LibsynAds.
现在就访问 libsynads.com，没错，就是 libsynads.com。
Go to libsynads.com that's libsynads.com today.
关于 Bayt 播客
Bayt 提供中文+原文双语音频和字幕，帮助你打破语言障碍，轻松听懂全球优质播客。