← Back to blog

Entrepreneurs · 9 min read · 16 Nov 2026

How to price a productized AI service

Pricing is where most first AI offers go wrong. Price by the hour and AI makes you cheaper every month. Price by gut feel and you either scare buyers off or work for less than your costs. This guide gives you a method: find your floor, anchor on the client's outcome, pick a structure, and write the scope so the price holds.

All numbers in this article are illustrative examples to show the method. They are not benchmarks, market rates or a prediction of what you will earn. Your prices depend on your market, your clients and your costs.

First, what you are pricing

A productized service has a fixed scope and a fixed price. The client buys a defined outcome, like "30 days of LinkedIn posts for a founder, drafted, edited and scheduled," not a block of your hours. AI is how you deliver it efficiently, not the thing you are selling.

This matters for pricing because it changes the question. You are not asking "what is my hourly rate?" You are asking "what is this outcome worth to this client, and what does it cost me to deliver reliably?" The price sits somewhere between those two numbers.

If you are still shaping the offer itself, start with what a productized AI service actually looks like, then come back here.

Cost-plus vs value-based pricing

There are two basic ways to set a price. You need both, but for different jobs.

Cost-plusValue-based
How it worksAdd up your costs, add a margin.Estimate what the outcome is worth to the client, charge a share of it.
What it is good forSetting your floor. The price you must never go below.Setting your actual price.
WeaknessAs AI makes delivery faster, the price keeps falling.Requires you to understand the client's business.
Question it answers"Can I afford to sell this?""Is this worth buying?"

Use cost-plus to find the line you cannot cross. Use value to decide where above that line to sit.

Step 1: Work out your cost floor

Your floor is what it costs you to deliver one unit of the service properly, including your own time. Most people forget half of these lines. Fill in a worksheet like this for one client for one month.

Cost lineWhat to includeExample (illustrative)
AI tool subscriptionsShare of Claude, ChatGPT or other plans used for this client.US$20
API and usage costsModel calls inside your workflows, per client.US$10
Other softwareScheduling tool, automation platform, forms, storage.US$25
Your delivery timeReview, editing, client calls. Hours × the rate you need.6 hrs × US$40 = US$240
Sales and admin timeInvoicing, onboarding, chasing approvals, spread per client.2 hrs × US$40 = US$80
Payment feesCard or transfer fees as a share of the price.about US$15
Buffer for reworkRevisions, a bad month, a model update that breaks a step.US$40
Total cost floorabout US$430 per month

In this made-up example, any price under about US$430 a month means the client is costing you money once your time is counted. That is the floor, not the price.

Two notes. First, count your time at the rate you need, not zero. "It only takes me a few hours" is how people end up busy and unpaid. Second, the AI tool costs are usually the smallest line. Your time and the admin around it are the real costs, which is why scope discipline matters so much (more below).

Step 2: Estimate the value to the client

Now look at it from the buyer's side. What does the outcome replace, save or create for them? You do not need precise numbers. You need a reasonable range you can talk about in a sales conversation.

Questions to ask the client, or to research before you talk to them:

Say, as an illustration, a founder currently spends about eight hours a month writing posts, or pays someone to do it. You can ask what those hours or that invoice cost them. That gives you a reference point grounded in their numbers, not yours.

A common rule of thumb is to price so the client clearly gets more value than they pay, often framed as the value being several times the price. It is a guide for judgment, not a formula.

Step 3: Choose a price structure

Productized AI services usually use one of four structures. Pick based on how the work behaves.

StructureHow it worksFits whenWatch out for
One-time fixed priceA single fee for a defined deliverable.A setup, an audit, a one-off build.No recurring income; you must keep selling.
Monthly retainerA fixed monthly fee for an ongoing outcome.Content, inbox handling, reporting, an AI agent you run.Scope creep month by month.
Setup fee plus retainerA one-time fee to build it, then a smaller monthly fee to run and maintain it.A custom automation or agent built for the client and then kept running.Clients questioning the monthly fee once it "just works."
Tiered packagesTwo or three fixed options with different volumes.When clients genuinely need different amounts.Too many tiers confuse buyers.

For setup plus retainer, make the monthly fee about something concrete: monitoring, fixing breakages when tools change, a monthly report, and a fixed number of small changes. "Maintenance" alone sounds like paying for nothing.

For tiers, three is usually the maximum. Many people find that the middle option is the one most buyers pick, so make it the one you most want to sell.

Step 4: Write scope boundaries that protect the price

A fixed price only works with a fixed scope. Without boundaries, a US$500 retainer slowly becomes a US$500 job that takes twice the hours. Every offer needs these written down before the client signs:

Clear scope is not unfriendly. Most clients prefer knowing exactly what they get.

Step 5: Price your pilot honestly

Your first clients are taking a risk on you. A pilot lowers that risk, but it should not mean working for free.

Options that tend to work better than free:

In exchange for the pilot terms, ask for something concrete: permission to describe the work, a short written comment if they are happy, or an introduction to one other potential client. Only use any of these with explicit permission.

Write the end date and the full price into the pilot agreement. Pilots without an end date tend to become permanent discounts.

Step 6: Plan how you will raise prices

Your first price is a starting point. You will learn your real costs and the real value to clients within a few months. Plan from day one how prices will change.

Step 7: Put it in a one-page promise

Before you send a proposal, write a one-page promise. This is the document that makes the price easy to say yes to. It should fit on a single page and answer these questions:

  1. Who it is for. One sentence describing the exact client.
  2. The problem. What it costs them today, in their words.
  3. The outcome. What they get, in numbers and dates.
  4. How it works. Three to five steps, from kickoff to delivery.
  5. What you need from them. Access, approvals, time.
  6. The price and structure. One clear number, what it covers, and how billing works.
  7. Scope limits. What is not included and how extras are priced.
  8. Your promise. What you guarantee, stated in terms you control, like delivery and quality standards. Never guarantee the client's revenue or results that depend on things outside your control.

If you cannot fit the offer on one page, the scope is probably not fixed yet.

Common pricing mistakes

MistakeWhy it hurtsBetter approach
Charging by the hourAI makes you faster, so your income per job falls.Charge for the outcome.
Pricing off tool costsTool costs are small; your time and risk are not.Use the full cost floor worksheet.
Vague scopeUnlimited work for a fixed fee.Numbers for every deliverable.
Too many optionsBuyers stall when choosing.One offer, or at most three tiers.
Promising results you do not controlDisputes and refunds.Promise delivery and quality, not their revenue.
Endless free pilotsTrains clients to expect free work.Short, paid pilots with an end date.

FAQ

Should I show my price on my website?

For a productized service, showing it usually helps. It filters out buyers who cannot afford it and makes the offer feel like a product. If you are still testing, you can show a "from" price.

Should I tell clients I use AI?

Be honest if asked, and consider saying it up front. Clients are buying an outcome and a person who stands behind it. Explain how you check quality. Some clients will also have policies about AI use with their data, so ask.

What if a client says AI makes this cheap, so it should be cheap?

Explain that they are paying for a reliable outcome, your judgment and your accountability, not keystrokes. Then point to what the alternatives cost them. If they still only want the cheapest option, they may not be your client.

How many clients do I need?

That depends entirely on your price, costs and goals. Work it backward: decide what you need, divide by your price minus your cost per client, and check whether you can deliver that many to a high standard. There is no typical answer.

Can I change the price after I launch?

Yes. Plan for it. Raise prices for new clients first and give existing clients written notice.

Price and pitch your offer, live

In the AI for Entrepreneurs track, session 2 designs a fixed-scope, fixed-price offer and its one-page promise, and session 5 covers pricing, proposals and a sales page. By session 8 you present your offer to the cohort. We cannot promise what any offer will earn; we can make sure you leave with one that is clearly scoped and priced. Starts January 2027.

Apply for January 2027Book a free 60-min session