5 Ways People Are Actually Making Money With AI

The secret most people overcomplicate

People making money with AI aren't creating money out of thin air. They're using AI to automate and speed up the things that were already making other people money.

Getting more reviews. Filling empty appointments. Answering the phone. Showing up when customers search. Running ads. Businesses have paid for every one of these for decades. AI just means one person can now deliver them, for a lot less work.

That's why these sell. You're not asking a business owner to believe in AI. You're offering something they already know they need.

The one rule: pick ONE play. Get one client. Then add the next. Trying to sell all five at once is how you end up selling none.

How to use the prompts

Every play below ends with a copy-paste prompt. Fill in the brackets and paste it into Claude (Cowork or Claude Code if you want it to actually build things on your computer) or ChatGPT.

The build prompts make the AI act like your technical co-founder. It asks you questions, explains every step in plain English, and builds it with you. You don't need to know how to code. You need to know how to answer questions about the business.

1. The Google review agent

Play 1 · Easiest to sell

After every visit, the customer gets a text with a direct link to leave a Google review. One reminder a few days later if they didn't. That's it. More reviews means a higher spot on Google Maps, and a higher spot on Google Maps means more customers walking in.

• Who pays for it
  Any local business that lives on Google Maps: dentists, salons, barbers, HVAC, plumbers, auto shops, restaurants, med spas. The easy tell: they have 30 reviews and the shop down the street has 400.
• How to sell it
  Look up their review count next to their top 2 competitors. Screenshot it. That screenshot is the pitch. Nobody has to imagine the problem, they can see it.

The stack

• Texting: Twilio, or the texting built into GoHighLevel if you want an all-in-one CRM
• Trigger: Whatever already knows a visit happened: the booking app, the POS, or a simple end-of-day CSV
• Glue: Zapier, Make, or n8n. Or skip all three and have Claude Code write a small script
• The link: The business's direct Google review link, from their Google Business Profile ("Ask for reviews" button)

Know this before you build

• Ask every customer, not just the happy ones. Google bans "review gating," meaning only sending the review link to people who said they were happy. It can get reviews removed.
• Never pay for reviews or offer a discount for one. Same reason.
• Only text people who gave their number and agreed to get texts. In the US, business texting also needs A2P 10DLC registration, which Twilio and GoHighLevel walk you through.

I want to build an automated Google review request system for a local business. Act as my technical co-founder and build it with me step by step. I'm not a developer, so explain each step in plain English before you do it.

The business: [business name, type, city, e.g. "Smile Dental, a 2-dentist practice in Crown Point, IN"]
How they book or check out customers: [e.g. "Jane App" / "Square POS" / "they export a CSV of the day's appointments"]
Their Google review link: [paste it, or tell me to find it later]
Texting tool: [Twilio / GoHighLevel / recommend one for me]

What the system should do:
1. When a visit is marked complete, wait [2 hours] and text the customer a short, friendly message in the business's voice with the direct review link.
2. If they haven't left a review after [3 days], send ONE reminder. Never more than that.
3. Never text anyone who replied STOP, and include opt-out language in the first message.
4. Send the review link to EVERY customer. Do not filter for happy customers first (Google calls that review gating and bans it).
5. Log every text sent (name, date, which message) in a simple Google Sheet so the owner can see it working.

Before you build anything:
- Ask me any questions you need answered.
- Tell me the cheapest stack that would work for a business this size and what it costs per month.
- Write 3 versions of the text message (under 160 characters each) so the owner can pick one.

Then build it, test it with my own phone number, and give me a 1-page "how this works" summary I can hand to the business owner.

2. The no-show killer

Play 2 · Easiest to prove ROI

Every empty appointment slot is money the business already lost. This agent texts reminders before the appointment, lets people confirm or reschedule by replying, and when someone cancels, it texts the waitlist to fill the spot. Fewer empty chairs, zero extra work for the front desk.

• Who pays for it
  Anything that runs on appointments: dentists, chiropractors, physical therapists, salons, barbers, med spas, tattoo shops, auto service, tutors.
• How to sell it
  Ask the owner one question: "How many no-shows did you have last week?" Multiply that by what one appointment is worth. Now they know what the problem costs them every month, and your price looks small next to it.

The stack

• Calendar: Their booking software. Check first: a lot of them already have basic reminders the business never switched on
• Texting: Twilio or GoHighLevel, with two-way texting so people can reply C to confirm or R to reschedule
• Waitlist: A Google Sheet of people who want an earlier slot is plenty to start
• Glue: Zapier, Make, n8n, or a script Claude Code writes for you

Know this before you build

• For anything medical or dental, keep health details out of the texts. "Reminder: your appointment at Smile Dental is Tue at 2pm, reply C to confirm" is fine. What the appointment is for is not.
• Same texting rules as play 1: opt-in, STOP works, A2P 10DLC registration in the US.
• Start with the boring version (reminders + confirm) before you build the waitlist fill. The reminders alone usually sell it.

I want to build an AI appointment reminder and no-show prevention system for a local business. Act as my technical co-founder and build it with me step by step. I'm not a developer, so explain each step in plain English before you do it.

The business: [business name, type, city]
Their booking software: [e.g. "Vagaro" / "Jane App" / "Google Calendar" / "I don't know yet"]
Texting tool: [Twilio / GoHighLevel / recommend one for me]
Average appointment value: [$ amount, or "help me estimate it"]

What the system should do:
1. Text a reminder [48 hours] and [2 hours] before every appointment. Short, friendly, in the business's voice.
2. Let people reply C to confirm or R to reschedule. If they reply R, send them the booking link and flag it for the front desk.
3. If someone cancels, text the first [3] people on the waitlist that a slot just opened, first to reply gets it.
4. Keep health and treatment details OUT of every message. Name, business, date and time only.
5. Respect STOP replies and include opt-out language.
6. Every Monday, send the owner a 3-line summary: appointments, confirmations, no-shows, and slots filled from the waitlist.

Before you build anything:
- First check whether their booking software already has reminders or a waitlist built in, and tell me what's missing.
- Ask me any questions you need answered.
- Tell me the cheapest stack that works and what it costs per month.

Then build it, test it end to end with my own phone number, and write a 1-page summary I can give the owner, including how to estimate what no-shows were costing them before.

3. The 24/7 voice agent

Play 3 · Highest perceived value

An AI receptionist that answers every call, day or night. It answers questions, books the appointment, and texts the owner a summary. Every call that goes to voicemail is a customer calling the next business on the list. This makes sure that never happens.

• Who pays for it
  Businesses where a missed call is a missed sale: home services (HVAC, plumbing, roofing), law firms, med spas, dental offices, auto shops. Especially anyone too busy on the job to pick up the phone.
• How to sell it
  Call the business after hours, or at lunch on a busy day. If it goes to voicemail, that's your opener: "I called you at 6:40pm on Tuesday and nobody picked up. Neither will a new customer."

The stack

• Voice: Retell AI (what the full build guide below uses). It handles the voice, the phone number, and the calendar booking
• Calendar: Cal.com or the business's own booking system
• Brain: A knowledge base: hours, services, prices, service area, FAQs. Claude writes it from their website in 5 minutes

Know this before you build

• I already wrote the full step-by-step build for this one, every screen and every setting. Use the full guide linked below to build it, and use the prompt here to write the agent's brain.
• Always give callers a way to reach a human. The agent should transfer or take a message for anything it can't handle.
• If the business records calls, the agent should say so at the start. Some states require everyone on the call to agree to being recorded.

The full AI Phone Receptionist build (13 steps, about an hour) →

I'm building an AI phone receptionist for a local business using Retell AI. Write everything the voice agent needs to sound like a great front desk person.

The business: [business name, type, city]
Their website: [URL]. Read it and pull everything useful.
Hours: [hours]
How appointments get booked: [e.g. "Cal.com link" / "they want a callback" / "transfer to the office line"]
Who to transfer urgent calls to: [name + number, or "take a message"]

Write me:
1. The agent's personality and greeting in 1 paragraph. Warm, quick, sounds local, not robotic. Say it's an AI assistant if someone asks.
2. A knowledge base: services, prices (only if they're on the website, otherwise say "we'll confirm on your visit"), service area, hours, parking, insurance or payment info, and the 15 questions callers most likely ask, each with a 1-2 sentence answer.
3. The booking flow, step by step: what info to collect (name, phone, reason, preferred time), in what order, and how to confirm it back.
4. Escalation rules: exactly when to transfer or take a message (angry caller, emergency, anything about billing disputes, anything it isn't sure about).
5. What to text the owner after every call: caller, reason, outcome, and anything urgent at the top.
6. 5 test calls I should make before going live, including 2 tricky ones, and what a good answer sounds like for each.

Never make up a price, policy, or service the business doesn't list. When unsure, the agent takes a message.

4. Get businesses mentioned in AI

Play 4 · Newest, least competition

People don't only Google "best dentist near me" anymore. They ask ChatGPT. And ChatGPT answers with 3 to 5 names. If a business isn't one of them, their competitor is. Your job: find out where the business stands in AI answers, then fix what's keeping them out.

• Who pays for it
  Any local business people research before they pick: dentists, lawyers, contractors, med spas, restaurants, gyms, real estate agents. This is so new that almost nobody is selling it yet, which is exactly why it's easy to sell.
• How to sell it
  Ask ChatGPT, Claude and Perplexity "best [their service] in [their city]." Screenshot the answers. If they're not in them and their competitor is, you just showed them a problem they didn't know they had.

The stack

• The audit: ChatGPT, Claude, Perplexity and Google's AI answers, asked the same questions a customer would ask
• The fixes: Their Google Business Profile, Bing Places, their website, review sites, and local "best of" lists
• The tracking: A Google Sheet where you re-run the same 10 questions every month and log who shows up

Know this before you build

• The big levers: a fully filled-out Google Business Profile, being listed on Bing Places too (not just Google), lots of recent reviews (hello, play 1), and a website that answers the exact questions people ask, in plain words, with the city named.
• The business's name, address and phone number should match exactly everywhere it's listed. Mismatches confuse AI the same way they confuse Google.
• Don't promise a #1 spot. Nobody controls what AI says. Promise the work, and show the before and after every month.

Act as an AI search visibility specialist. I want to find out why a local business does or doesn't show up when people ask AI tools for a recommendation, and build a plan to fix it.

The business: [business name, type]
City / area they serve: [city]
Their website: [URL]
Their top 3 competitors (if I know them): [names, or "find them"]

Step 1, the audit. Search the web and act like a customer. Run these and tell me who gets recommended:
- "best [service] in [city]"
- "[service] near [city] open on Saturday"
- "most trusted [service] in [city]"
- "affordable [service] in [city]"
- 6 more questions a real customer would ask before picking a [service].
For each one: which businesses got named, whether my business showed up, and what sources the answer seems to be pulling from.

Step 2, the gap. Compare my business to whoever shows up most. Check their Google Business Profile (categories, photos, review count and recency, Q&A), Bing Places, website (does it answer these questions in plain words with the city named?), and which review sites and "best of" lists mention them.

Step 3, the plan. Give me the 10 fixes most likely to get the business mentioned, ranked by impact and effort, with exactly what to do for each one. Then write:
- 5 FAQ answers for their website, in plain language, each naming the city.
- An improved Google Business Profile description (under 750 characters).
- A one-page before/after report template I can send the owner every month.

5. Run ads that pay for themselves

Play 5 · Biggest monthly retainer

This used to be the hard one, because good ads took a designer, a copywriter and a week. AI now makes the creative and writes the copy in minutes. So your whole job is the part that matters: testing what works and cutting what doesn't. Bring in more than the ads cost and businesses will pay you every single month.

• Who pays for it
  Businesses with a clear offer and a way to take money or bookings online: med spas, gyms, home services, restaurants with a promo, local retail, ecommerce brands.
• How to sell it
  Look up the business in the Meta Ad Library (it's free and public). Most local businesses are running nothing, or one tired ad from two years ago. Show up with 5 finished ad mockups for their business. That's a hard thing to say no to.

The stack

• Research: Meta Ad Library: see what every competitor is running right now, for free
• Creative: ChatGPT or Higgsfield for images and short video, Claude for the copy and the angles
• Running it: Meta Ads Manager. The business pays Meta directly for the ad spend. You charge for managing it
• Tracking: The Meta pixel on their site, or a simple count of calls and bookings from the ad

Know this before you build

• Have the business pay ad spend on their own card, straight to Meta. You charge a separate monthly fee for running it. Never front their ad spend.
• Agree on what "working" means before you launch (booked calls, sales, cost per lead) so the conversation in month 2 is about numbers, not feelings.
• Test lots of different creative, not lots of tiny tweaks. Five very different angles beats one ad in five colors.

The AI Money Stack (the 5 tools behind these plays, with what to charge) →

Act as a direct-response ad strategist for local businesses. I want to create a first round of Meta ads for a business and pitch them on running ads.

The business: [business name, type, city]
Their website: [URL]. Read it.
Their main offer, if I know it: [e.g. "$99 new patient exam" / "free estimate" / "I don't know yet"]

Step 1, research. Look at what this business and its top 3 local competitors are saying on their websites and reviews. Tell me:
- The 5 biggest reasons customers pick a business like this (pull real phrases from reviews where you can).
- The 3 biggest fears or objections that stop people from booking.

Step 2, angles. Write 5 very different ad angles (not 5 versions of one idea). For each: the angle in one line, who it's for, and why it should work.

Step 3, the ads. For each angle write:
- A headline under 8 words for the image itself.
- The primary text (under 125 characters before the "see more" cut).
- A short image or video description I can paste into ChatGPT or Higgsfield to generate the creative. Real-looking, local, not stock-photo looking.

Step 4, the pitch. Write a short, friendly message I can send the owner with these mockups attached. No hype, no guarantees, just: here's what I'd test first and why.

Only use claims you can back up from their website or reviews. If you'd need a number or a promise from the owner, put [CONFIRM WITH OWNER] instead of making it up.

Which one should you start with?

• Play 1, the review agent

  If you want your first sale this month:

  Easiest to explain, cheapest to build, and the competitor screenshot sells it for you.

• Play 2, the no-show killer

  If you want the clearest money story:

  The owner already knows what an empty slot costs. You just do the multiplication.

• Play 3, the voice agent

  If you want the "wow" demo:

  Let the owner call it themselves. That 60-second call does more than any pitch.

• Play 4, AI mentions

  If you want to be early:

  Almost nobody is selling it yet, and every business owner asks AI questions themselves.

• Play 5, ads

  If you want the biggest monthly check:

  The highest retainer of the five, but you have to bring results to keep it.

How to land your first client

Same 5 moves, no matter which play you pick.

• 1. Pick one play and one type of business
  Not "local businesses." Dentists in your town. HVAC companies in your county. Narrow makes everything easier: the outreach, the build, the case study.
• 2. Build it for free once
  For a friend's business, your barber, your old boss. You need one real setup you can point at. The free one is how you learn and how you get your proof.
• 3. Lead with what you found, not what you sell
  The review count next to the competitor. The voicemail at 6:40pm. The ChatGPT answer they're missing from. Show the problem first. The offer is just the fix.
• 4. Charge monthly
  Every one of these runs every day, so price it that way. Set it up, keep it running, send a short report every month showing it working.
• 5. Stack the plays
  Every one of these makes the others work better. Reviews help them show up in AI. The voice agent books the calls the ads create. Your first client for one play is your easiest sale for the next.

Help me write a short outreach message to a local business owner about [the play, e.g. "getting more Google reviews"].

The business: [name, type, city]
What I found: [the specific proof, e.g. "they have 31 Google reviews, their top competitor has 412" / "I called at 6:40pm Tuesday and got voicemail" / "ChatGPT recommends 4 dentists in their city and they aren't one of them"]
How I'm reaching out: [DM / email / walking in / text]

Rules:
- Under 80 words.
- Open with what I found, specific and true. No flattery, no "I hope this finds you well."
- One sentence on what I'd do about it.
- End with a low-pressure question, like "Want me to show you what that would look like for you?"
- Sound like a real person from their town, not an agency. No buzzwords, no guarantees.

Write 3 versions: one direct, one friendly, one that leads with a question.