How to Price Your AI Services (And Stop Undercharging)

A guy in the AI Avengers community posted his first AI project last year. He built a lead qualification bot for a mortgage broker. The bot processed 30 form fills per day, flagged the hot ones, and sent a personalized follow-up within 5 minutes of submission.

He charged $800 for the project. One-time fee. Took him 12 hours to build.

The mortgage broker closed 4 extra deals in the first quarter after installing it. At $3,000 average commission, that is $12,000 in additional revenue - from a bot he paid $800 for.

He undercharged by roughly $15,000.

The Hourly Rate Trap

Most people pricing AI services default to hourly. It feels safe. Transparent. Fair. It is also a terrible way to price work that compounds in value over time.

Hourly pricing caps your income at your hours. It also punishes efficiency - the faster you get at building AI systems, the less you make per project. That is backwards.

Value-based pricing ties your fee to the outcome you produce. If a bot generates $5,000/month in additional revenue for a client, charging $500/month is 10% of the value you create. That is a deal for them and sustainable income for you.

How to Calculate Value Before You Quote

The formula is straightforward. Identify the metric your AI system improves. Find the dollar value of that metric. Charge 10-20% of the annual value as your fee.

Lead qualification example: Client gets 200 leads/month, closes 5% ($10,000 average deal). Current team qualifies leads manually - 3 hours/day, misses 30% due to slow response. Your bot captures 30 more leads/month that would have gone cold. At 5% close rate that is 1.5 extra deals. At $10K average = $15,000/month in value. Charge $1,500-3,000/month.

Time savings example: Owner spends 15 hours/week on tasks your system automates. Owner's time is worth $200/hr (what they could be billing or the cost of a hire). 15 hrs x $200 x 4 weeks = $12,000/month in value. Charge $1,200-2,400/month.

You do not need precision here. You need a number that anchors the conversation. "This system should generate roughly $5,000-8,000/month in additional revenue based on your current close rates. We charge $750/month." That is a 15:1 ROI conversation.

The Pricing Structure That Works

Setup fee + monthly retainer. This is the standard that sustainable AI service businesses use. It accounts for the build work upfront and the ongoing maintenance and optimization.

For small businesses (local service companies, individual practitioners, small agencies): $2,000-5,000 setup fee + $500-1,000/month retainer. This range is accessible for most business owners and still profitable for you.

For mid-market clients (established businesses, larger agencies, professional services firms): $5,000-10,000 setup + $1,000-2,000/month. These clients have bigger budgets and bigger problems. The ROI calculation is easier to justify.

The retainer is the important part. It locks in recurring revenue, keeps you engaged with the client, and creates opportunities to expand the engagement. One client on a $1,000/month retainer for 12 months is $12,000 - more than most one-time projects.

How to Present ROI in a Proposal

The proposal is where most people lose deals. Not because the price is too high - because the value is not clear. Here is the structure that closes:

Section 1: The problem they told you about. In their words. "You mentioned that your team spends 3 hours per day on lead follow-up and still misses same-day responses on 40% of new inquiries." This shows you listened. It also anchors the conversation on their pain, not your service.

Section 2: The solution. What you will build, what it does, how it connects to their existing tools. Keep it plain language - no technical jargon unless they are technical.

Section 3: ROI projection. Conservative estimate. "Based on your current conversion rates, capturing 30 additional qualified leads per month should generate 1-2 additional deals. At your average deal value of $4,500, that is $4,500-9,000/month in additional revenue."

Section 4: Timeline. Most clients want to know when they will see results. 2-3 weeks to build, 1-2 weeks testing, live in 30 days. Specific is better than vague.

Section 5: Price. Setup fee plus monthly. Make the monthly feel obvious given the ROI section. "The investment is $3,500 setup and $750/month. Based on the conservative projection above, this pays for itself in the first additional deal."

Common Objections and How to Handle Them

"Can we start lower and increase as I see results?" This is a negotiation move, not a real objection. The answer is: "The setup fee is fixed because it covers the build. We could reduce the monthly to $500 and revisit after 90 days when you have seen the results." You protect the upfront and show confidence in your work.

"I want to see it work before I commit to ongoing." Give them 30 days at the full setup fee. If they do not want to continue after 30 days, you still got paid for the build. Most clients who see it working keep going.

"That's more than I expected." Acknowledge it, then return to the ROI. "I hear you - it is a meaningful investment. The reason we price this way is because we are pricing on the outcome, not the hours. If this generates $5,000/month, you are making money from day one."

What Not to Do

Do not send a quote without a discovery call. You need their numbers to build the ROI case. A blind quote is always too high or too low.

Do not reduce the setup fee to close a deal. It signals desperation and undervalues the work. Reduce the monthly retainer if you need to move on price - that is easier to increase later when results are proven.

Do not take on projects where you cannot calculate value. If a client cannot tell you what a lead is worth or how much a task costs them, you cannot build an ROI case and you will end up working cheap.

The Numbers That Build a Business

5 clients at $1,000/month retainer = $5,000/month recurring. Add 1-2 new setup projects per month at $3,000-5,000 each = $8,000-15,000/month total. That is a real business.

Getting there from zero takes 3-6 months if you are focused. The first client is the hardest - usually someone in your network who trusts you enough to be your proof of concept. Charge them fairly, get documented results, use those results in every proposal after.

AI Avengers Lab has a community of people building this exact business model. We share proposals, pricing examples, and client results. If you want to skip the trial and error, that is the fastest path. Join at aiavengers.team.

Frequently Asked Questions

What if I don't have case studies yet?

Offer your first client a reduced setup fee in exchange for a documented case study. $1,500-2,000 setup instead of $3,000-5,000, with the condition that they share before/after numbers after 90 days. That case study will make every subsequent deal easier.

Should I charge differently for different industries?

Yes. Healthcare, legal, and financial services have higher deal values and higher risk - charge more. Local service businesses (HVAC, plumbing, restaurants) have thinner margins - the ROI framework still applies but deal values are lower. Calibrate to the industry's economics.

How do I handle clients who want to own the system outright?

Charge a one-time fee 3-5x your normal setup fee for full ownership transfer. Include 90 days of support at a flat rate. After that, charge hourly if they need changes. Many clients say they want ownership but continue the retainer when they see the ongoing value.

What if the ROI doesn't materialize as projected?

Always use conservative projections and present them as estimates, not guarantees. If results fall short, investigate why - often it is a data quality issue or a workflow the client did not follow. Fix it proactively rather than waiting for the client to raise it.

Related reading from this series

This post is part of the Claude Side Business playbook. The full series covers every step with concrete workflows, pricing, and lessons from running my own business on Claude.

For more playbooks, visit the AI Avengers home page or join the AI Avengers Skool community to put these into practice with weekly office hours.