The argument in one line.
Local businesses with expensive, already-painful problems are the fastest path to $10K with AI, and Claude can handle every step from niche research to signed proposal without the consultant writing a single line of code.
Read if. Skip if.
READ IF YOU ARE…
- A freelancer or generalist who wants to start earning from AI skills but has no existing audience or product.
- A career-switcher willing to have direct conversations with small business owners rather than build consumer apps.
- Someone who already knows a specific industry even tangentially and wants to turn that familiarity into a consulting edge.
- A builder stuck in learning mode who wants a concrete one-week action plan to get a first paying client.
SKIP IF…
- You are building a consumer SaaS product -- this is a services model, not a product model.
- You want passive income mechanics; this approach requires active client conversations.
- You need a deep technical build guide -- this covers the business model and workflow, not implementation specifics.
The full version, fast.
The Expensive Problem Filter requires three things at once: the problem costs the business money every month, the business has real revenue to pay you, and you can prove the return in dollar terms fast. Two problems pass every time -- leads slipping through the cracks and the dead CRM database. Once niche and problem are locked, Claude handles the entire workflow: industry research, outreach copy, discovery call transcript analysis, build plan, and final proposal. The consultant's job is to have the conversation and steer the output, not to write code.
Free for members
Chat with this breakdown — free.
Sign in and you get 23 free chat messages on us — ask for the hook, quote a framework, find the exact transcript moment, generate a markdown action plan. Bring your own key when you want unlimited.
Create a free account →Where the time goes.
00:00 – 00:22
01 · Hook and credibility anchor
Opens with the zero-starting-point premise and establishes personal credibility via the insulation-company-to-$100K arc.
00:22 – 01:05
02 · Video promise
Maps out everything covered: which problem, how to find clients, what to say, what to charge, how Claude does the heavy lifting.
01:05 – 01:54
03 · The two traps
Names the learning loop and building-before-selling as the two patterns that keep beginners broke.
01:54 – 02:34
04 · The reframe: local retainer model
Reframes the goal from selling products to holding retainers with local businesses. Real deal: $5K setup plus $500/mo.
02:34 – 03:12
05 · Community CTA
Pitches the free Skool community and 4-week challenge mid-video.
03:12 – 03:49
06 · The Expensive Problem Filter
Introduces the three-criteria filter: monthly pain the owner feels, business has money, provable return.
03:49 – 04:34
07 · Problem 1: Leads slipping through the cracks
Voice agent case study: $5K plus $500/mo, saved $1,600/mo vs receptionist, recovered $50K in revenue.
04:34 – 05:31
08 · Problem 2: Dead database
HVAC 5,000-lead example: 5% reactivation x $3K job = $750K. Pricing $2K-$8K or $3K plus rev-share.
05:31 – 06:12
09 · Why boring niches win
HVAC, plumbing, insulation, dentists, roofers: recession-proof, real revenue, underserved by AI consultants.
06:12 – 07:32
10 · Claude as research engine
Use Claude to research the niche, rank expensive problems, write outreach copy in the prospect's language. Find 10 businesses on Google Maps or local Facebook groups.
07:32 – 08:50
11 · Discovery call and Claude build plan
Do not pitch -- ask questions. Record the call, feed transcript to Claude, get build plan and pricing from the client's own words.
08:50 – 09:30
12 · Claude builds and proposes
Claude builds in plain English with no code. Claude writes the proposal with scope, price, terms. Send, sign, collect.
09:30 – 10:19
13 · Zero case studies objection
Build one proof asset. Frame as outcome numbers not feature lists. One result beats a wall of testimonials.
10:19 – 11:14
14 · Mindset close
The real skill is finding an expensive problem and having the conversation. Not code, not tools.
11:14 – 11:57
15 · Weekly assignment and CTA
Pick one problem, one niche, find 10 businesses, send the message, book one call. Subscribe ask.
Lines worth screenshotting.
- The two traps that keep AI beginners broke are the learning loop and building before anyone agrees to pay for it.
- Local boring-niche businesses have real revenue, daily operational pain, and almost no AI consultants competing for them.
- A voice agent that catches missed calls and follows up in seconds can recover $50,000 in revenue for a single client, making a $5,000 setup fee feel trivial.
- A dead CRM list of 5,000 old HVAC leads at 5% reactivation and $3,000 per job is worth $750,000 -- the campaign fee stops feeling scary framed against that number.
- The discovery call is where the build plan is hiding: record it, feed the transcript to Claude, and it outputs what to build, how to build it, and what to charge.
- One proof asset documented as an outcome statement beats a wall of testimonials -- people buy the result, not the resume.
- The outcome frame that closes deals: I helped an insulation company book 14 estimates in 30 days, save 2 hours a day, and add $118,000 in annual revenue.
- Putting the building in front of the selling is the root error -- money is always on the other side of that order.
- Claude can write the outreach message, the discovery questions, the build plan, and the full proposal -- the only non-delegatable job is having the conversation.
- The real skill is not technical: it is finding a business with an expensive problem and being willing to have the conversation about it.
Find the expensive problem before you build anything.
WHAT TO LEARN
Most early AI consulting failures share a single root cause -- they prioritize building over selling, and discover too late that nobody agreed to pay for what they built.
- The learning loop is a comfort trap: consuming one more tutorial feels productive but delays the income conversation indefinitely, and readiness never arrives on its own.
- Local service businesses -- HVAC, plumbing, insulation, dentists, roofers -- have real monthly revenue, daily operational pain, and almost no AI competitors targeting them.
- The Expensive Problem Filter has three non-negotiable gates: the problem costs money every month the owner already feels it, the business has cash to pay you, and you can quantify the return -- all three must be true simultaneously.
- Two problem types pass the filter consistently: leads that slip through to voicemail and competitors, and dormant CRM databases full of past customers who could be reactivated with a structured campaign.
- Discovery calls are information extraction sessions, not pitches -- ask what happens when a lead comes in, how fast someone calls them back, and what one job is worth.
- Recording the discovery call and feeding the transcript to an AI is a legitimate method for generating a scoped build plan and pricing rationale grounded in the client's own words.
- A proof asset framed as outcome numbers -- estimates booked, hours saved, revenue added -- does more selling work than a feature list or a portfolio of logos.
- The gap between someone who earns from AI and someone who only watches videos about it is not technical skill; it is whether they send the first slightly awkward outreach message.
Terms worth knowing.
- Expensive Problem Filter
- A three-criteria test: the problem costs the business money every month and the owner already feels it, the business has revenue to pay you, and you can quantify the return quickly enough to justify your fee.
- Dead Database
- A CRM or contact list full of old leads and past customers that a business has stopped following up with, which can be reactivated through an automated outreach campaign.
- Learning Loop
- The trap of consuming tutorials and testing new tools indefinitely instead of sending outreach, driven by the belief that readiness arrives once you know enough -- which it never does.
- Proof Asset
- A single documented result described as outcome numbers rather than feature lists, used as an entire portfolio when starting from zero.
- Retainer
- A recurring monthly payment from a client for ongoing maintenance or optimization of an AI system, providing predictable income beyond the initial setup fee.
Things they pointed at.
Lines you could clip.
01:45
“You're putting the building in front of the selling when the money is on the other side of that order.”
03:20
“When a problem is already bleeding them, you're not convincing them to spend. You're just showing how to stop the bleeding.”
05:28
“Every AI creator out there is chasing that shiny consumer app, which means the plumber and the insulation company down the street are wide open. And that gap is your whole opportunity.”
08:35
“I don't know how to code. I never learned. And I build everything in plain English by handing Claude that plan and letting Claude Code do the actual work.”
10:40
“The real skill was finding a business with an expensive problem and being willing to have the conversation about it.”
11:43
“The difference between those two people is honestly not talent. It's whether you close this video and go send something or close this video and open another tutorial.”
Word for word.
Read-along
Don't just watch it. Burn it in.
See every word as it's spoken — crank it to 2× and still catch all of it. The same dual-channel trick behind Amazon's Kindle + Audible.
metaphoranalogystory
If I had to start completely over today with no audience, no clients, no money, and no list to fall back on, and I had to make my first $10,000 with Claude, here's exactly what I would do.
And I'm not talking in theory here because a year and a half ago, I was making $50 a year doing marketing for an insulation company. I'd barely even touched chat GPT. Months later, I'd crossed a $100 as an AI consultant doing the exact thing that I'm about to walk you through.
I'm gonna show you the one problem that I would go after first because it pays the most and honestly it's the easiest one to prove. And then how I'd find the business, what I'd actually say to get them on a call, what I'd charge, and how I'd use Claude to do all the heavy lifting at every single step. From the research all the way down to the actual build.
So that by the end of this video, you know the exact move that you'll be making this week. And the part that almost nobody gets right is which problem to pick in the first place. Because most people pick the one that's almost impossible to get paid for.
So let me save you a few months of spinning your wheels. Here's the thing. If you search how to make money with AI right now, you're going to get buried under the same advice and most of it keeps you broke.
The first trap is the learning loop and I say this with love because I've lived in it. Just watching one more tutorial, learning your fifth tool this month, and telling yourself that you'll start reaching out the moment you understand how all of this really works. And you will never feel ready because there's always a newer tool and a longer video.
So you just keep learning and you never earn a dollar. The second trap is building something that's polished before anybody has agreed to pay you for it. You sit there for weeks building a beautiful product that nobody has asked for, and then you take it out into the world and find out the hard way that nobody wants it.
Both of these have the same root problem, which is that you're when the money is on the other side of that order. So let me reframe this whole thing for you because once this clicked for me, everything changed. You need a couple of local businesses, each one sitting on a single expensive problem, who will pay you every month to make that problem go away.
Let me put real numbers on it so it stops feeling abstract. One of the first deals that I ever closed was $5,000 plus $500 a month to keep it running, and that client stayed because businesses don't churn the same way that some app subscribers do.
So one client like that is already half of your first $10 and they keep paying long after the setup, which means you're not starting from zero again every thirty days, the way you wouldn't be if you were chasing tiny one off sales. I help builders and freelancers transition to AI consultants who close 4 and 5 figure projects and retainers and build something that actually lasts.
And you if wanna land your first or next AI client, you can take part in the four week challenge inside my free school community of like minded people, where you will find a free course that's honestly better than most paid courses out there, and you'll get direct one on one access to me for whatever questions you might have.
The link is down below. Now, let's get into it. So which problem do you actually go after?
Because this is the part I promised you, and it's the part that decides whether you make money or not. I run every opportunity through what I call the expensive problem filter. And a problem only passes if it does three things at once.
So let me walk you through them. The first one is that it's costing the business money every single month and they can feel it. Not a nice to have, not a someday thing, but money walking out the door right now in a way that owner is already annoyed about it.
Because when a problem is , you're not convincing them to spend. You're just showing how to , and that's a much easier conversation. The second is that it's a local business that actually has the money to pay you, which is why I love the boring ones.
HVAC, plumbing, insulation, dentists, roofers, the kind of business that everybody needs and nobody posts about.
Most of them are doing real revenue. They're not luxuries that disappear during a downturn, and they can comfortably afford a few thousand dollars every month if you're making them more than that. And the third is that you can prove the return fast because a problem that you can put a dollar figure on is a problem that you can charge a lot to solve.
Let me give you the two that pass this filter every time. The first is lead slipping through the cracks, and this is the one I would lead with. I built a voice agent for a business once, charged them $5,000 up front plus $500 a month, and it saved them about $1,600 a month versus the receptionist they were paying.
And then it caught all of the calls that used to go to voicemail and then straight to their competitor. Two months later, the owner told me that it had brought in at least $50,000 in revenue just because the leads were getting followed up with in seconds instead of sitting in a voice mailbox for three days.
That's an expensive problem, and you can see exactly why she kept paying. The second one is the dead database, and this one feels like it's printing money for the client.
Almost every one of these businesses has a list of old leads and past customers just sitting in a CRM doing absolutely nothing. So picture an HVAC company with around 5,000 old leads. And even if only 5% come back, that's 250 jobs.
And at a $3,000 average job, that's $750,000 in revenue on a list that they'd already given up on. I charge somewhere between 2 and $8,000 for a campaign like that depending on the size of the list.
Or I'll do 3,000 upfront plus a cut of whatever it brings in. And when you frame it as found money against the number that big, the price stops being scary. And I know the boring niches don't sound exciting, but that's exactly why they work.
So stay with me. These businesses have money because everyone needs their AC fixed or their pipes working whether or not the economy is good. They have real daily pain because half of them are still running on paper forms and spreadsheets and a person that they're paying full time to do that work that software should be handling.
And almost nobody is actually building for them because every AI creator out there is , which means the plumber and the insulation company down the street are wide open. And that gap is your whole opportunity. And I know this one personally because I literally worked for an insulation company.
So when I started, I went after the exact businesses that I already understood. And you can do the same with whatever world that you already know, even just a little. So here's the actual path from nothing to your first signed client.
And I want you to watch how Claude does the heavy lifting at almost every step because this is where the whole thing goes from sounding hard to feeling almost unfair. You start by picking one problem in one niche and going deep instead of being a little useful to everyone. So let's say you're gonna be the person that fix a slow lead follow-up for home service companies.
The first thing I would do is open up Claude and have it research that world for me. I'd literally ask it to break down the HVAC industry, the pain points that cost these owners the most money, the actual language and slang they use, what a single job is worth, and where these people hang out online.
And in about a minute, you can get back the kind of understanding that used to take months of working in that industry. So by the time you actually talk to one of these owners, you already sound like you live in their world. Then I'd have Claude take that same research and rank the most expensive problems for me.
So instead of guessing, I'm asking it what two or three problems in this niche a business would happily pay $10 or more to fix and roughly what each is costing them each month. Now I'm not pulling a problem out of thin air because Claude just handed me the ones with real money attached. Once I know the problem I'm solving, I'd have Claude write the outreach for me in their language so that it doesn't sound like a template.
And it would be something just as simple as, hey, I help HVAC companies turn the leads that slipped through into book jobs without hiring anyone new. Would you be open to fifteen minutes this week? Then I find 10 of these businesses on Google Maps or in a couple local Facebook groups, which is plenty to start.
And then I send that message to all 10. When one of them gets on a call, you do not pitch. You ask questions and let them tell you what's broken and what it's costing them because the whole build plan is hiding inside of their answers.
So you're asking things like, walk me through what happens when a lead comes in. How fast does somebody actually call them back? And what's one job worth to you?
And here's the move that I love because with their okay, you record that call. And then you take the transcript and drop that entire thing into Claude and have it pull out the built plan and the pricing for you. At that point, the client is basically telling Claude what to build since you're just feeding it their own words and out comes a plan that lays out what to build, how to build it, and what to charge based on what they just told you the problem is costing them.
Then comes the part that everyone thinks is the hard part, which is actually building the thing and you can do that in Claude too. I don't know how to code. I never learned.
And I build everything in plain English by handing Claude that plan and letting Claude code do the actual work. So whether it's the campaign that goes back through their dead database or the system that follows up with new leads in seconds, Claude is the one putting together while you just steer it.
And when it's time to close, I'd have Claude write the proposal as well. So you end up with a clean professional document with a scope, the price, the terms, the kind of thing that makes you look like a real professional company your very first week.
You send that over, they sign, and now you've made your first real money using Claude without you having to write a single line of code yourself. And I know what some of you are thinking which is, Lindsay, that's great, but why would anyone pay me when I have zero case studies? Here's what I did and what I would do again from scratch.
I build one proof asset instead of waiting for the perfect testimonial to magically appear. I went and got a result for one business and I documented it, and that became the only proof that I needed. And the way that you talk about it matters way more than you think because nobody cares that you built a chatbot with 15 integrations.
That's features and it makes people's eyes glaze over. What lands is the outcome. So the way I would say it is, I helped an insulation company book 14 estimates in thirty days.
Save two hours a day and add a $118,000 in annual revenue. Here's how I did it.
That's it. And if even that feels like work, you hand clawed the result and have it write the little proof video script for you in two minutes. So there goes that excuse too.
One short outcome like that does more than a wall of testimonials because people buy the result, not the resume. So if you're sitting there telling yourself that you're not technical enough for this, I need you to really hear me because this is the thing that changed my life.
A year and a half ago, I was making $50 at that insulation company, checking my bank account anytime that I bought groceries, and I had barely even used ChatGPT. I wasn't some genius developer. What I figured out was that the building was not actually the hard part because the tools get easier every single month, and they still are.
And the real skill was finding a business with an expensive problem and being willing to have the conversation about it. So the whole thing is to find a painful expensive problem in a niche that you can get to know, charge real money to make it disappear, and then make the return so obvious that saying no would be the strange choice.
That's how I went from $5,000 projects and $500 a month retainers to over a $100,000 in nine months.
And none of it required me to be a coder first. So here's what I actually want you to do this week. And I don't mean save the video and feel productive.
I mean actually do it. Pick one problem and one niche, and I'd point you towards slow lead follow-up in a home service business because it's the easiest one to prove. Then find 10 of those businesses near you, send that message to every single one, and aim to book just one conversation where you shut up and just let them tell you what it's costing them.
That's the entire assignment. It's free and you can start today. Because a year from now, you're either gonna be the person that watched a stack of videos just like this one and never sent a single message or you're gonna be the person that sent that first slightly awkward one, stumbled through a couple of calls, and ended up with a client paying them every single month.
And the difference between those two people is honestly not talent. It's whether you close this video and go send something or close this video and open another tutorial. So if this helped, go ahead and like this video and subscribe if you actually wanna learn how to make money with this stuff because I put out a new one every week.
Drop your questions in the comments and I will get to them. Thanks for watching and I will see you in the next video.
The Hook
The bait, then the rug-pull.
A year and a half before this video, the host was earning $50 a year doing marketing for an insulation company and had barely touched ChatGPT. The hook lands because it is a real before -- not aspirational positioning, but a verifiable low point -- which makes the $100K-in-nine-months claim that follows feel earned rather than inflated.
Named ideas worth stealing.
03:12list
The Expensive Problem Filter
- It costs the business money every single month and they can feel it
- It is a local business that actually has the money to pay you
- You can prove the return fast -- put a dollar figure on it
A three-gate qualifying test for deciding which problems to pursue. All three must be true simultaneously.
Steal forany consulting or agency pitch process
06:12list
The Claude End-to-End Workflow
- Claude researches the niche: pain points, language, job value, where owners hang out
- Claude ranks the most expensive problems in that niche
- Claude writes outreach copy in the prospect's language
- Record discovery call, feed transcript to Claude, Claude outputs build plan and pricing
- Claude builds the solution in plain English
- Claude writes the proposal with scope, price, and terms
A six-step sequence where Claude handles every production task and the consultant handles only the client-facing conversation.
Steal forany AI-assisted services business; template for systematizing client delivery with LLMs
09:30concept
The One Proof Asset
Get one result, document it as outcome numbers not features, and use that single case as your entire portfolio. Format: I helped X do Y in Z days.
Steal foranyone starting a services business without a portfolio
How they asked for the click.
VERBAL ASK
02:34product
“You can take part in the four week challenge inside my free school community of like minded people, where you will find a free course that's honestly better than most paid courses out there.”
Dropped mid-video at the natural seam after the reframe and before the core content. Secondary subscribe CTA at the very end. Pitch is confident -- she says free and better than most paid courses without hedging.
MENTIONED ON CAMERA
FROM THE DESCRIPTION
AFFILIATECommission earned if you click.
Visual structure at a glance.
