Corey Gannon charges $999 for a 45-minute conversation.
No code. No audience. No startup capital. He doesn't build anything.
He sits down with a small business owner, finds their biggest time drains, and prescribes off-the-shelf AI tools that fix them. Like a doctor writing a prescription.
Then 50 to 60% of those clients ask him to implement the fixes. That upsell path took him to $8K MRR in 10 days with his consulting retainer, which works out to over $1,000 an hour.
Greg Isenberg had him on the pod, and Corey gave away the entire playbook. I watched the full hour so you don't have to. Here it is, with all the real numbers.
The offer: $999 or your money back
Corey's clients are small business owners with 2 to 20 employees, doing $500K to $5M a year in revenue.
The pitch is simple. Give me 45 minutes. I'll interview you, pull out your bottlenecks, and prescribe 3 to 7 off-the-shelf AI tools you can implement immediately to reclaim 5 to 10 hours per week.
And the kicker: if he can't find at least 5 hours a week of savings, you get 100% of your money back.
Worst case, you lose 45 minutes. Best case, you get your evenings back. His average client reclaims 7 hours a week.
He's sold about 15 of these assessments in 2026 so far. $999 sounds small, but that's the point. It's the foot in the door. The lifetime value of an assessment client ends up at $3K, $5K, $10K or more once the upsells kick in.
Why does this work at all? Because business owners are drowning. They're working 40, 50, 60 hours a week, they're terrified their competitors are getting ahead with AI, and by Corey's estimate maybe 5% of small businesses use any AI tool beyond prompting ChatGPT.
You're not selling technology. You're selling hours back.
The four-phase delivery system
This is where most people expect complexity. There isn't any.
Phase 1: the discovery call. A 45-minute Zoom with an AI note taker running (he uses Fathom, but Otter or Fireflies work too). He only asks questions. "Walk me through your day yesterday." "What tasks do you dread?" "Where does work pile up?"
His favorite: "If you could wave a magic wand and delete any process in your business, what would it be?"
The most common answer is email. Every time.
The rule on this call: never prescribe. Corey says he has to bite his tongue because he usually knows the fix within minutes. Probing only. The pitch comes later.
Phase 2: AI analysis. He feeds the call transcript to Claude with a skill that pulls out pain points and researches tools that fix each one. Then he QA's the output himself. If Claude prescribes Salesforce to a 4-person landscaping crew, he subs in a CRM that actually fits a small business.
His two research shortcuts: futurepedia.io and theresanaiforthat.com. Both are giant AI tool directories, tagged by industry.
On your first 2 or 3 assessments, the AI gets you 60 to 70% of the way there. Feed your finished reports back into the skill so it learns what good looks like, and by assessment five or six it's nearly copy-paste.
Phase 3: the report. This is the client-facing deliverable, and he's iterated on it 12 times. One goal: stupid simple. His line stuck with me: a confused mind doesn't buy, doesn't implement, doesn't get ROI, and doesn't upsell.
The report has six pieces:
- An executive summary (main pain points, hours reclaimed per week)
- An effort vs. impact matrix (the report only targets high-impact, low-effort "quick wins")
- The recommendations: pain point → tool → cost → setup time → hours saved
- A 4-day quick start plan (one 10-minute action per day)
- A "what comes after quick wins" slide that tees up the upsell
- A financial impact slide: weekly time returned × the owner's hourly rate, minus tool costs
The average tool stack he prescribes costs about $60 a month. For 7 hours a week back. The net ROI is always four figures a month, sometimes five.
He open-sourced the whole template at audittemplate.ai. Literally plug and play.
Phase 4: the review call. 30 minutes, screen sharing the report, walking through each recommendation. Then three closing questions. Which of these is most urgent? Do you want to do this yourself, or do you want my help? What's your timeline?
50 to 60% of clients answer: "Can you just do it for me?"
That's the entire business. Your client pays you $999 to uncover reasons to pay you more.
The upsell menu (where the real money is)
Once the assessment surfaces the problems, Corey has six things to sell next:
- Process redesign: $3,000 to $3,500. One client, an e-commerce seller, was burning 10 hours a week monitoring Amazon ad spend and wanted it automated. Corey's answer: fix the broken 16-step process first, or the automation falls apart on top of it.
- Automation builds: simple Zapier, Make or n8n flows. He charged a wedding venue $1,500 to automate their Asana project setup for every new client.
- Knowledge systems: a business broker was getting 400 to 500 emails per listing, all asking the same five questions. Corey trained a custom GPT on the listing's marketing package. Emails answered per deal now: zero.
- Custom workflows: Claude skills for a company's proprietary processes, around $3K, with an optional maintenance retainer since processes change.
- Full implementation: any combination of the above, packaged around $8,000.
- AI concierge: his favorite. More on that below.
One closing trick worth stealing: when a client hesitates on a $5K implementation, he credits their $999 assessment toward it. Except he's already marked the project up an extra $1,000. Same revenue, and the client feels like they got a discount.
(Technically he makes a dollar more. Greg called it out on the pod and they both cracked up.)
The AI concierge: $1,000 an hour, capped at six clients
This is the recurring revenue layer, and it's almost embarrassingly simple.
The offer: two 45-minute Zoom calls a month. The client shares their screen. Corey helps them build Claude skills for their actual workflows. Between calls they get unlimited Voxer access with a 12-business-hour response time.
When he came up with it, he had no idea if anyone would buy. He pitched the first six people he talked to and closed five. He also raised the price on every yes: $1,200 a month, then $1,500, then $1,800, then $2,000.
$8K MRR in the first 10 days.
Do the math on $1,500 a month for two 45-minute calls. That's $1,000 an hour. Corey says he's not a fan of trading time for money, but at that rate he'll make an exception. (Same.)
And the Voxer access that sounds like a support nightmare? Of his five clients, three have messaged him. Once each. In two and a half months. Huge perceived value, near-zero workload.
On the calls he runs one loop on every workflow, which he calls AOA: Audit, Optimize, Automate. Show me how you do it today. Cut the 13 steps down to 7. Turn it into a Claude skill. Rinse and repeat for as long as they're a client.
It's the same systems thinking Brett used to automate an entire podcast producer job with Claude skills. The difference is Corey gets paid $1,000 an hour to do it with someone watching.
Two retention details that matter. Before the first call, an onboarding form asks: "What would make this engagement feel like a win in 90 days?" Four of his five clients hit their 90-day win in about 60 days. And every engagement lives in a Notion hub with a running inventory of everything shipped, emailed after every call. That's what makes someone keep paying $1,800 a month without blinking.
Seven ways he finds clients with no audience
The most common objection: "Sounds great, but I have no followers and no money for ads." Corey's answer is seven methods that cost nothing:
- Host a local AI meetup. Barter with a coworking space (free conference room in exchange for foot traffic), give a 20-minute talk on how to use Claude, collect contacts at the door, follow up within 24 hours. He partnered with a well-connected realtor for his first one in January and got a client from it.
- Door knocking. A listener DM'd him: 30 local businesses, 5 meetings, 2 clients. Out of 150,000 people who heard Corey describe the tactic, one person did it. It worked.
- LinkedIn DMs, but local and human. No pitch in message one. His opener to the local HVAC owner: "I probably live down the street from you."
- Free mini audits for your network. Text a business owner you know: 15 minutes, free, I'll show you one way AI helps your business. The full $999 assessment is the natural next step.
- Agency partners. Your accountant, your insurance agent, marketing agencies. Their clients are asking them AI questions they can't answer. Offer a referral fee and follow up every 2 to 3 weeks.
- AI office hours. Sit in a coworking space one hour a week as the free AI guy. A guy in Corey's community asked 10 spaces, got 2 yeses, door-knocked local restaurants for $175 in gift cards to raffle, and pulled 11 attendees and 2 follow-up calls from his first session.
- Post your wins. Build in public, even when the win is just "someone texted me a ChatGPT question and I nailed the answer."
Notice the pattern: every method targets people within driving distance. Being "the AI guy in Charlotte" beats being one of 10,000 AI consultants on the internet.
What I'd steal from this playbook
You might not want to run AI audits for landscapers. I don't either. But four things transfer to almost any service or SaaS business:
→ Charge for discovery. A paid audit flips the dynamic: the client pays you to find reasons to hire you again. Free proposals do the same work for $0 and less respect.
→ Make the deliverable stupid simple. Corey rebuilt his report 12 times, mostly by deleting. Confused clients don't implement, and clients who don't implement don't renew.
→ Ladder the pricing. $999 assessment → $1,500 to $3,500 project → $1,500 to $2,000 a month retainer. Each step is a small yes that makes the next one easier.
→ Constrain supply, honestly. Five clients, capped at six, and he says so on every sales call. Real scarcity closes deals that discounts can't.
The bigger thread: AI finally gives a one-person business real leverage. It's the same idea as the loop engineering playbook Elie Steinbock broke down: build a system once, let it compound. Corey just wrapped the system in a service instead of a SaaS.
FAQ
How much can you make with an AI audit business?
Corey charges $999 per assessment and has sold about 15 in 2026. Since 50 to 60% of clients buy implementation afterward ($1,500 to $8,000 per project), the lifetime value of one client runs $3K to $10K+. His consulting retainer alone hit $8K MRR within 10 days of launching.
Do you need to know how to code to start this?
No. The core offer is prescribing off-the-shelf tools that already exist, not building anything. You research fixes with Claude and directories like futurepedia.io and theresanaiforthat.com. You do need to genuinely understand the AI tool landscape, which comes from doing the first few assessments.
How do you get the first client with no audience?
Start with the free mini audit: text one business owner you already know and offer 15 minutes of free AI recommendations. It's the lowest-friction method on Corey's list, and the full assessment is the obvious upsell. If you want clients this week, he'd tell you to go knock on doors.
What is an AI concierge service?
A done-with-you retainer. The client pays $1,200 to $2,000 a month for two 45-minute working sessions where you build AI workflows together on their screen, plus async access between calls. You're charging for the gap between "I know AI matters" and "I'll never set this up alone."
Get the full playbook every week
Stories like this are exactly why I run the Profitable Founder Podcast. Every week I sit down with bootstrapped founders making $100K to $10M a year and pull out the numbers, the pricing, and the playbooks they actually used.
