Build a $10K/month AI business with zero code and zero cold outreach

You don't need to do cold outreach or know how to code to start making money selling AI services.

I built my AI services business past $10K a month in under a year with a three-stage offer ladder and a list of 10 people I already knew.

This article breaks down the full model: the pricing, the $10K math, the warm-network sales process, and the three mistakes that kill your margin.

Before we dive in, I bundled my 4 best resources into one easy download:

1. The Mini AI Assessment Playbook
2. AI Tools Assessment Report Template
3. AI Concierge Playbook
4. 8 Common AI Services Objections (and how to destroy them)

You can grab all 4 resources (free) here.

Stage 1: Run a free mini AI assessment

The free mini assessment is a 15-minute conversation with a business owner.

Your job is to find one painful bottleneck and prescribe one practical fix.

Ask questions like:

• Where does work tend to pile up?
• Which manual task keeps stealing time from you or your team?
• What have you tried to automate before that didn't work?
• If one process got easier this month, which one would matter most?

You'll probably uncover several problems. Rank the top three, then focus on the best opportunity.

The prescription can be an off-the-shelf AI tool or one AI workflow the owner could build. If the owner spends five hours a week sorting email, you might recommend SaneBox as a simple starting point. The recommendation is the deliverable.

Don't install it for free.

That boundary matters. You're giving the owner a useful plan, not donating an implementation project.

After the first call, schedule a five- to 10-minute follow-up. Walk through the three opportunities you found, explain the one you'd fix first, and give the owner a simple next step.

Then make two asks:

1. Ask for a testimonial or referral if the assessment was useful.
2. Offer a paid assessment to find three to seven more opportunities.

Only run this free version for your first one to three prospects. Its job is to build confidence, collect proof, and create a natural transition into paid work.

Stage 2: Sell the deeper diagnosis

The paid AI tools assessment uses the same basic process, but goes much deeper.

Instead of a 15-minute interview, spend 45 to 60 minutes mapping the owner's most important workflows. Instead of prescribing one fix, identify three to seven AI opportunities.

Score each opportunity on two dimensions:

• Impact: How much money, time, or capacity could this improve?
• Effort: How hard will it be to implement and maintain?

Start with the high-impact, low-effort opportunities. In most cases, those are existing AI tools the business can adopt without a custom build.

The assessment should also document the assumptions behind the ROI.

Ask the owner how much each bottleneck costs in time every week. Then ask what an hour of their time is worth.

Say a business owner spends five hours a week in email and values their time at $250 an hour. That makes email a $1,250 weekly time cost. If a tool cuts the task from five hours to one, the estimated time value returned is $1,000 a week.

That's an estimate, not audited savings. The important part is that the owner supplied both inputs. You're not making up a dramatic ROI number after the call.

Package the findings in a professional report that includes:

• The workflows discussed
• The top bottlenecks
• Three to seven recommended tools or workflows
• An effort-versus-impact ranking
• The assumptions behind each ROI estimate
• A prioritized action plan

(you can download the exact template I use for my reports for free at audittemplate.ai)

Then schedule a 30-minute follow-up to present it. Walk through each problem, your recommended fix, and why it earned its place on the roadmap.

I started pricing these assessments at $500. I later sold them for $1,000, which was high enough for clients to take the work seriously and still easy for qualified owners to approve. I currently sell them for as much as $2,000.

No testimonials yet? Start at $1,000. That's what worked for me. Raise the price as your proof and delivery improve.

The assessment makes money on its own, but its bigger job is exposing the work that belongs in a recurring AI concierge engagement.

Stage 3: Turn the roadmap into a monthly retainer

AI concierge is done-with-you consulting for owners who know they need AI but don't know how to apply it to their business.

My standard engagement includes:

• Two 45-minute strategy calls each month
• Ongoing voice-message access through Voxer
• One scoped workflow improvement at a time
• Help building reusable Claude Skills inside Claude Cowork
• Documentation and team training for the new process

The work follows a framework I call AOA: audit, optimize, automate.

Audit the process

Ask the client to share their screen and show you how the work gets done today. Capture each step without trying to fix it on the spot.

If customer onboarding takes 30 minutes, watch the full sequence. Note where information comes from, which tools the employee opens, what decisions they make, and where the process slows down.

Optimize it before adding AI

Most manual processes contain unnecessary steps.

If the current onboarding workflow has 14 steps and only needs 11, remove the waste first. Automating a bloated process only makes the bad version run faster.

Automate the cleaner version

Once the workflow is clear, turn the optimized process into a Claude Skill. During the call, guide the client through the build, the prompts, and the setup.

This is done with the client, not handed over as a black box. Adoption goes up because the owner understands what the skill does and how it fits into the business.

After the skill works, use Claude to help document the new SOP and train the rest of the team.

Then move to the next workflow.

The list rarely gets shorter for long. Clients cross off one problem, then notice two more opportunities because they finally understand what AI can do inside their operation. That gives the engagement room to continue as long as the work keeps producing value.

The math behind $10K a month

The paid assessments generate useful project revenue. The recurring concierge clients are what make the model predictable.

The monthly math can work several ways:

• 10 clients at $1,000 per month = $10,000
• 7 clients at about $1,430 per month = about $10,000
• 5 clients at $2,000 per month = $10,000

My current concierge clients pay between $1,200 and $2,000 a month. I plan to quote my next new engagement at $2,500.

The delivery model is intentionally light. Two 45-minute calls equal 1.5 hours of scheduled call time each month. At $1,500 per month, that's $1,000 per scheduled call hour before Voxer support, prep, and admin. At $2,000, it's about $1,333 per scheduled call hour before that extra work.

Those numbers only hold if you protect the scope.

Start with people who already trust you

You don't need cold email or a large following to find the first client.

Take out your phone and write down 10 people you know who run small businesses. Friends from the gym or church, parents of your kids' friends, former colleagues, past clients, local business owners you see regularly.

Text three of them and offer the free mini assessment.

The rough math: three free assessments produce one paid assessment, and two paid assessments produce one concierge client. This is a simple funnel to work toward based on my experience with the offer.

If you contact all 10 warm prospects, a reasonable target is five or six free assessments, two or three paid assessments, and one concierge client.

The strongest prospects are owner-led service businesses with repetitive work, visible bottlenecks, urgency, and enough budget to act. My concierge clients have included an insurance agency, an info-product business, an eCom business, a financial advisor, and a fintech company.

Many of the best-fit businesses I see are doing roughly $3 million to $10 million or more in annual revenue.

They already know the manual work is a problem. They don't need a lecture about why AI matters. They need someone to show them where to start and stay close enough to help them use it.

Three ways this model breaks

The offer is simple, but it's easy to ruin the economics.

1. You turn the free assessment into free implementation

One bottleneck. One prescription. Stop at the roadmap.

If you build the workflow during the free engagement, you remove the reason to buy the paid next step and teach the client that your implementation time has no value.

2. Two calls quietly become unlimited calls

My cadence is two 45-minute calls each month. I respond to Voxer messages within 12 business hours, but I don't add unscheduled Zoom calls for free.

Set that expectation before the engagement begins. If a client wants more calls, charge for the added access.

3. You build a different technology stack for every client

Trying to master Claude, ChatGPT, Codex, Zapier, and a pile of niche tools at the same time will crush your margin.

I standardize the core fulfillment process around Claude Cowork and Claude Skills. The skills can be adapted if the preferred AI platform changes, but I don't relearn the entire delivery stack for every account.

Standardization is what turns consulting into a business you can repeat.

The first 10 actions I would take

If I were starting this model from zero today, I would:

1. Make a list of 10 business owners in my network.
2. Invite three of them to a free 15-minute mini assessment.
3. Use the same interview questions every time.
4. Rank the top three bottlenecks for each owner.
5. Recommend one tool or workflow, without implementing it.
6. Ask for a testimonial, referral, and paid follow-up.
7. Price the deeper assessment between $500 and $2,000 based on my proof.
8. Turn the assessment plan into a recurring concierge proposal.
9. Use AOA to improve one workflow per call.
10. Raise my concierge price by $250 after each accepted proposal.

That's the model.

The free assessment opens the door. The paid assessment proves you can diagnose the business. The concierge retainer pays you to help the client act on the plan one process at a time.

You don't need to become an engineer. You need to ask good questions, recommend useful tools, improve a process before automating it, and keep the client moving.

That's enough to build a real AI services business.

If you want the deeper playbooks for the mini assessment, paid assessment, AI concierge offer, and common sales objections, download the free AI services resource library here.

If you want the complete business in a box, including scripts, templates, skills, a 60-day roadmap, and community support, join AI Operator Academy.

Be sure to follow me @coreyganim for daily posts on growing an AI services business.