How a regular guy built a one-person AI agency to $50k contracts

You don’t need to be a programmer, influencer, or marketing guru to build a real AI business. One solo founder went from zero technical background to landing a $50,000 consulting contract and a $3,500 setup + $1,500/month client in just a few months—powered mostly by Claude and simple no-code tools.

This is a breakdown of what he actually did, how he priced it, and the mindset that kept him going before the money showed up.

From retirement to AI agency owner

After retiring from his previous work in mid-2025, Phil decided to start an AI automation agency. He wasn’t a developer, had no coding history, and no formal background in marketing or social media.

What he did have was curiosity, a willingness to learn by doing, and access to modern AI tools—especially Claude and Go High Level (a popular CRM and automation platform).

His core idea: use AI to solve real problems for small businesses, then charge for the implementation and ongoing management.

Starting simple: voice agents with no coding

Phil’s first move was to build AI voice agents inside Go High Level. These were basic phone or chat agents that could answer questions, capture leads, and handle simple customer interactions.

He:

• Built a few simple agents in Go High Level
• Showed them to friends and local roofing companies
• Used those demos to start conversations about what AI could do in their businesses

He wasn’t trying to build a perfect product. He just needed something real enough to start meaningful business conversations.

How a casual pitch turned into a $50,000 contract

One of those early conversations was with a friend who ran an investment firm managing around $1.4 billion across nine offices. The owner was brilliant with finance, but had no idea how to practically use AI in his business.

Phil didn’t pitch a big package right away. Instead, he:

• Talked with him multiple times over several months
• Asked questions about how the firm worked
• Built trust by explaining what AI could realistically help with

Eventually, the owner asked a simple question: if he hired Phil on a part-time retainer for six months, what would it cost?

Together they landed on:

• A six-month consulting agreement
• Roughly 20–25 hours per week
• Total contract value: $50,000 (paid in six monthly checks)

Phil’s main deliverable: AI-powered content creation and automation for the firm’s marketing and social media, plus other internal automations as needed.

Building a content engine with Claude

For the investment firm, Phil used Claude to build a powerful content creation system based on the owner’s existing material.

The firm had:

• 22,000+ chunks of past content (radio shows, scripts, notes, etc.)
• 570+ existing commercials and marketing pieces

Phil and Claude turned this into a custom knowledge base that could:

• Generate new content in the owner’s exact voice
• Draft social posts, scripts, and marketing messages
• Keep everything consistent with the firm’s brand and expertise

Phil’s pitch was simple: if the owner didn’t like the output, he essentially didn’t like his own voice—because the system was built entirely on his content.

A second win: $3,500 setup + $1,500/month in 12 hours of work

Another friend ran a nonprofit and a chiropractic practice. Phil first helped the nonprofit by building a grant-writing assistant in Claude called “Grant Craft.”

This app could:

• Find grants that matched the nonprofit’s mission
• Analyze each grant’s requirements
• Draft applications based on content about the nonprofit
• Let a human review and finalize before submission

Phil built this for free as proof of concept, with a performance-based pricing model in mind: $1,500 per grant application, $2,000/month retainer, and a cap of 8% of any grant awarded.

That project built trust—and led directly to a bigger opportunity in the chiropractor’s main business.

Product launch automation for a local clinic

The chiropractor bought a new FDA-approved body-toning laser and wanted to launch it as a service. He had an established practice and 7,500 existing patients, but no functioning CRM and no idea how to run a launch.

Phil stepped in and offered to handle the entire launch and ongoing management. His pricing:

• $3,500 for initial setup
• $1,500 per month for at least six months
• $75 per new client per month beyond the first 10 clients
• The client pays all Go High Level CRM costs
• An additional $500/month ad budget for paid marketing

What Phil actually built:

• A content creation app focused on the specific laser device (Zerona Z8)
• An automated bot in Go High Level to message all 7,500 existing patients and answer questions
• A sign-up form and workflow for consultations and treatments
• Ongoing social media and campaign management for the new service

He estimates the initial build took about 12 hours—even though he was learning parts of Go High Level as he went. After that, the monthly work dropped to roughly 1.5 hours, while the retainer stayed at $1,500/month.

How he uses Claude as his “silent business partner”

Phil openly says he doesn’t know how to code. He learned Claude by just talking to it and iterating. Over time, Claude became his behind-the-scenes partner for almost everything:

• Brainstorming solutions after client meetings
• Turning messy notes into clear project plans
• Drafting contracts and proposals
• Working out pricing structures
• Designing simple internal tools and apps

He even uses Claude to prep questions before meeting a client in an unfamiliar industry, and then again afterward to design solutions based on what he heard.

If you want to go deeper on this style of building, it’s very similar to using Claude to launch a one-person business, as described in this guide to launching a one-person e‑commerce business with Claude.

Recording every client conversation

One of Phil’s most valuable habits is simple: he records every client meeting.

He originally used a wearable recording pendant, but a smartphone voice memo works just as well. His process:

• Hit record before the meeting starts
• Focus fully on listening instead of taking notes
• Upload the audio for transcription
• Paste the transcript into Claude

From there, he and Claude:

• Identify the client’s real pain points
• Brainstorm multiple solution options
• Draft a proposal and pricing
• Prepare answers for likely questions or objections

This turns every conversation into raw material for custom AI solutions and makes it much easier to deliver something valuable after the meeting.

Finding clients: start with the people you already know

Phil’s early mistake was trying to “sell AI” instead of listening. He would talk too much in meetings, explaining all the things AI could do, instead of letting business owners share their problems.

Once he fixed that, his client acquisition strategy was straightforward:

1. Go through every contact in his phone (over 1,000 people).
2. Make a list of anyone who owned or worked in a business.
3. Send a simple text: “I’ve started an AI automation business. I’d love to sit down and see if there’s any way I can help your business.”
4. Aim for conversations, not hard sells.

This led to 7–10 sit-downs with business owners and, eventually, to his first major contracts. Every single paying client so far has come from:

• Direct friends
• Friends of friends
• People he already had some connection with

He hasn’t even fully tapped referrals yet—and he admits he hasn’t been asking existing clients to introduce him to others, even though that’s an obvious next step.

What he actually asks in those meetings

Phil doesn’t show up with a long checklist. His main goal is to get the business owner talking. Typical moves:

• Ask open-ended questions about their business and daily operations
• Let them vent about what’s frustrating or time-consuming
• Listen for anything repetitive, manual, or content-heavy

If he doesn’t understand the industry, he asks Claude beforehand to help him come up with a few smart starter questions. But during the meeting, he mostly listens.

Most owners, he’s found, already know three things:

• They should be using AI somehow
• They don’t know where to start
• They definitely don’t know how to implement it

That’s the gap he fills.

Simple, practical pricing strategies

Phil’s pricing isn’t complicated, but it is thoughtful and tailored to each situation. Some patterns:

1. Retainers for ongoing work
For the investment firm, he priced based on weekly hours and value, not just tasks. The owner actually suggested $50,000 for six months after they discussed scope.

2. Setup fee + monthly management
For the chiropractor’s laser launch, he charged:

• A one-time build fee ($3,500)
• A monthly management fee ($1,500 for six months)
• Performance-based upside ($75 per client beyond 10)

3. Performance-based and capped pricing
For the grant-writing tool, he structured it so the nonprofit only pays once grants are awarded, with a cap of 8% of the grant amount. That keeps incentives aligned and makes the offer easier to say yes to.

In all cases, he uses Claude to help think through pricing models, compare options, and draft clear proposals.

Why content creation is the “new business sign”

Phil keeps coming back to content creation, even though he has no formal background in it. The reason is simple: most local businesses have almost no consistent content presence.

He treats content like a storefront sign in a busy town:

• Years ago, a business could survive without a sign—everyone just knew where it was.
• Now, with more competition, you at least need a basic sign so people know you exist.

AI-powered content doesn’t have to go viral to be valuable. It just needs to:

• Keep the business visible and relevant
• Reflect the owner’s real voice and expertise
• Show up consistently where customers are already looking

With tools like Claude and connected platforms (for example, Beehive’s new MCP integration that lets AI see real newsletter data), it’s getting easier to build content systems that are both personalized and data-driven. If you’re interested in this kind of AI-agent workflow, you may also like this guide to getting started with AI agents in your business.

Mindset: staying motivated before the money arrives

There was a gap of months between Phil starting his AI agency and landing his first big contract. During that time, he had lots of meetings and no revenue to show for it.

His way of staying motivated:

• Do the hard thing first. He leans on the “eat the frog” idea: tackle the task you’re dreading (like outreach or sales calls) before anything else.
• Learn by action, not by theory. Instead of waiting to “know enough” about AI, he learned by taking on real problems and figuring them out with Claude as he went.
• Remember you know more than they do. Most business owners know almost nothing about AI. You don’t need to be an expert—just far enough ahead to guide them and know where to get answers.
• Talk to yourself like you’d talk to your kids. When he fails or gets discouraged, he reminds himself not to shame or belittle himself. Failure is an event, not an identity.

Above all, he believes we’re living in the best time in history to start a business. AI lets you build things you’ve never built before, for people who desperately need help but don’t know where to start.

How to copy this playbook in your own way

If you want to follow a similar path, here’s a simple starting roadmap based on Phil’s story:

1. Pick a simple tool stack. For example, Claude + a no-code CRM/automation platform like Go High Level or similar.
2. Choose one simple offer. Voice agents, content creation, basic automations, or AI-powered business assessments are all good entry points.
3. Mine your contacts. Go through your phone, LinkedIn, or email. Make a list of every business owner or decision-maker you know. Text or message them to set up a conversation.
4. Record every meeting. Use your phone’s voice memo app. Transcribe and feed it to Claude to design solutions and proposals afterward.
5. Price for value, not hours. Use setup fees, monthly retainers, and performance-based upside where it makes sense.
6. Let AI be your co-founder. Use Claude (or your preferred model) to help with brainstorming, planning, pricing, contracts, and even learning unfamiliar tools.

You don’t need to be a coder. You don’t need a big audience. You just need the willingness to talk to people, listen for pain points, and let AI help you build solutions you’ve never built before.