How to Make Money With AI in 2026: 10 Real Ways, Ranked

How to Make Money With AI: The Short Answer

Building and selling your own AI app, on the App Store or the web with a builder like Superapp, is still one of the best ways to make money with AI, because it is the only common path where the income can keep growing without your hours growing with it. It is also the hardest path on this list: most apps never reach 1,000 USD a month, and the work that decides success (finding users, keeping them, and paying for AI usage) starts after the app is built. The other nine ways below, from AI automation services to freelancing and content, pay sooner but are capped by your time. Every way below comes with real people describing their results in their own words, including the ones who made nothing.

Quick answer: how do you make money with AI? Superapp and similar builders make the highest-ceiling path practical: build an AI app that solves one specific problem and charge a subscription. AI apps earn more than average, with RevenueCat reporting revenue per install above 0.63 USD after 60 days, about twice the overall median, but results are extremely uneven and most apps earn very little. Faster, lower-ceiling options are selling AI automation to local businesses, freelancing with AI tools, creating content with AI help, and selling AI-assisted digital products. Avoid anything that promises passive AI income for an upfront fee; the FTC has sued schemes like that.

The 10 Ways to Make Money With AI, Ranked

Superapp's own path, building an app, is ranked first for earning potential, not for ease. Read the effort and time columns before choosing.

How to read this: the ways at the top have the highest ceilings and the slowest, least certain starts. The ways in the middle trade your time for money more directly. The ways at the bottom are easy to start, which is exactly why they are crowded.

1. Build and Sell an AI App on the App Store or the Web

Superapp exists for this path: you describe an app, Superapp builds it as a native iPhone, iPad, Mac, or Apple Watch app or as a web app, and you sell it with a subscription. Of all the ways to make money with AI, this is the one most likely to become a real business, and the one most likely to quietly earn nothing. Both are true at once, so this section covers both in depth.

Why an AI app is still one of the best ways to make money with AI

The case rests on four things that no other path on this list combines.

It is recurring revenue that does not need your hours. A freelancer or consultant earns when they work. An app earns while you sleep, from users you have never spoken to. Subscriptions mean one good month can carry into the next.

AI apps earn more per user than most apps. RevenueCat's State of Subscription Apps 2025 found that "Most AI apps see revenue per install above $0.63 after 60 days," about twice the overall median of $0.31 (RevenueCat). People pay for apps that do real work for them, and AI lets a small app do a lot of work.

The distribution and payment rails already exist. Apple says over 850 million people visit the App Store every week, and that 70% of visitors use search to discover apps (Apple Ads). Apple handles payments, refunds, and tax, and developers earning up to 1 million USD a year can join the App Store Small Business Program at a "reduced commission rate of 15%" (Apple Developer). On the web, Stripe charges "2.9% + 30¢ per successful transaction for domestic cards" (Stripe), so you keep more per sale but find every customer yourself.

Building no longer requires a team. Two years ago, a native iPhone app meant learning Swift or paying a developer. Now AI builders write and compile the code, which moves the bottleneck from building to everything around it.

Real AI apps that make money

Cal AI was launched in May 2024 by Zach Yadegari with three co-founders. According to Fortune, it brings in about 30 million USD a year, a figure Fortune says it confirmed with financial records. Yadegari's explanation of the growth is not about the AI: "We started making TikToks, we started partnering with influencers, and then we had pretty explosive growth."

Photo AI is Pieter Levels' web app for AI photoshoots. It launched in February 2023, and in August 2024 Levels announced it had passed 100,000 USD in monthly revenue (Indie Hackers). Levels had been building and launching products in public for years before this one took off.

Base44, a web app that builds other apps, was about six months old when Wix bought it for 80 million USD in cash in June 2025 (TechCrunch). Founder Maor Shlomo had kept it bootstrapped and reported 189,000 USD in profit in May, "after covering high LLM token costs." By the time of the sale it had eight employees.

These are the exceptions, and they are listed so you can see what the top looks like, not what to expect. The next section is about what the typical outcome looks like.

Smaller AI apps, in their founders' own words

Superapp users are more likely to land somewhere between the giant outcomes above and zero, so these are the stories worth studying most closely: small teams and solo founders with real numbers, told in their own words.

Florian Vates, MonAi: from about 300 USD to 12,000 USD a month through one creator. Florian builds native iOS apps. His main app, MonAi, is a minimalist expense tracker that uses AI to make entering expenses easier, and his portfolio earns roughly 40,000 to 60,000 USD a month (Indie Hackers). "Currently, I mainly work on MonAi because it generates 90% of my income," he says. The turning point was not a feature: "Most of our growth occurred after I collaborated with Charlie Alvarez, a content creator from Colombia." Charlie's story-driven videos in Spanish took MonAi from about 300 USD to 12,000 USD in monthly recurring revenue within two months, and he is now Florian's business partner. Florian's lesson: "If I did it again, I would definitely find a partner sooner."

Lorik Morina, RotateProduct: about 6,800 USD a month with 20 USD a day in ads. Lorik built an AI app that turns one product photo into a rotating 3D video for online sellers. At week 21 it was earning 6,772 USD a month, with growth split roughly evenly between paid ads and organic (MRR Story). "I set up Meta (Facebook/Instagram) ads spending just $20 a day," he says. His rule for whether ads are working is simple: "When you spend $20/day and it brings back more than $20 in subscription value, you have a money-printing machine." And on waiting for luck: "I didn't wait for a magical viral moment."

An iOS developer known as caonema: from 200 USD to 3,000 USD a month. In a candid post, this developer described what did and did not grow their iOS app (Indie Hackers). Posting on forums and social media and running Google Ads mostly brought Android users who could not use the iOS-only app: "I quickly find this not work." Switching to App Store optimization and Apple Search Ads did: "More traffic indeed bring more download and revenue." Moving the paywall to after onboarding increased long-term subscribers. And the part nobody plans for: "Refund are sucks." The developer estimated refund requests had been costing about 900 USD a month before they took steps to manage them.

Max Artemov: from a failed app to a 30-app portfolio earning about 22,000 USD a month. Max's first app, a calorie counter, got nowhere after more than five years of work (Indie Hackers). He changed his approach completely: each new app is built around one core feature and one App Store keyword, and he credits ASO with raising impressions, downloads, and revenue by about 50%. "The most important advice I can offer is this: Don't fear shipping," he says. "Only after I see real usage and positive feedback do I add more features." His working method now: "I just build, release, and let the product sink or float based on real user feedback."

Why it is not trivial: the honest part

Building the app is now the easy part. These are the parts that decide whether it makes money, and most people underestimate every one of them.

Outcomes are extremely uneven. RevenueCat found that the top 5% of newly launched apps earn 8,880 USD after two years, more than 400 times the bottom 25%, who earn 19 USD or less (RevenueCat). The same research found that "most apps fail to garner $1,000 in monthly revenue within their first two years" (TechRepublic). An app is not a salary. It is closer to a lottery ticket you can improve your odds on, with skill and persistence.

AI is not a differentiator anymore. RevenueCat's own summary of its AI findings carries the warning: "But AI alone won't drive success, differentiation does." Thousands of apps wrap the same AI models. If your app is a chat box with a theme, a competitor can copy it in a weekend, and the AI provider can add your feature to its own app for free.

Distribution is the actual job. Look at the three stories above: Cal AI credits TikTok and influencers, and Levels had an audience built over years. None of them credit the code. A new app has no reviews, no rankings, and no audience. Getting the first thousand users usually takes more work than building the first version, and it never really stops.

App Review is a real gate. Apple reviewed 9,100,620 app submissions in 2025 and rejected 2,093,244 of them, including 415,532 in the design category that covers copycats, minimum functionality, and spam (Apple Transparency Report 2025). Guideline 4.3(b) says: "Don't submit apps that are indistinguishable from what's already widely available," and names categories like dating, wallpaper, simple timers, and fortune telling where Apple will "not accept new submissions unless they offer a meaningfully different or improved experience" (App Review Guidelines). AI apps carry extra rules: apps that create AI content need content filtering, reporting, and blocking (guideline 1.2), and you must "clearly disclose where personal data will be shared with third parties, including with third-party AI, and obtain explicit permission before doing so" (guideline 5.1.2(i)).

AI costs come out of every user. A normal app costs almost nothing per extra user. An AI app pays for every request it sends to a model. Base44's founder described its profit as earned "after covering high LLM token costs." If you price a subscription without knowing what a heavy user costs you, your best customers can lose you money. On Apple devices, Apple's on-device Foundation Models can handle some tasks with no per-request bill, which is one reason native apps have an advantage here.

Subscribers leave. RevenueCat found that "Nearly 30% of annual subscriptions are canceled in the first month." Every month, you need new users just to stay level. An app that people try once and forget will never grow, no matter how good its marketing is.

It is a business, not a project. Once people pay, you have support emails, refunds, reviews to answer, bugs on devices you do not own, privacy policies, tax forms, and operating system updates that break things each autumn. None of it is hard on its own. All of it together is a part-time job before the app pays like one.

Your output may not be protected the way you expect. The US Copyright Office concluded in January 2025 that purely AI-generated material is not copyrightable and that prompting alone does not give enough control to make you the author (Copyright Alliance summary). Your app's value has to come from what you design, choose, and build around the AI, not from the raw output.

The apps that did not work, in their founders' words

Superapp can build an app in days, which makes it more important, not less, to hear from founders whose apps did not earn. Their posts are rarer than success stories, and more useful.

DanialPG built an AI tool for web agencies and, after 40 days of SEO and community work, had zero paying strangers. He published five blog posts, built Reddit karma, launched on Product Hunt, and posted on Indie Hackers. Search impressions grew from zero to 1,050, which produced 6 clicks (Indie Hackers). "The people who signed up are people who already knew me," he wrote, and he was honest about the uncertainty: "I genuinely don't know if I need more patience or a different approach entirely." Forty days is short for SEO, which is exactly the point: the slow channels are slow.

Max Artemov's first app took more than five years and went nowhere. His later success with a 30-app portfolio came from changing the method, not from persisting with the first idea. He describes being too attached to the original product to see that users did not need most of its features.

Guy Powell's warning about AI wrappers. In an essay titled "Your AI Product Is Not A Real Business," Powell argues that most AI startups put a thin interface over someone else's model and have nothing defensible (Indie Hackers). "AI is not your product," he writes. "A demo is not a product." It is the same point RevenueCat makes with data: AI alone does not drive success; differentiation does.

The math of an AI subscription

Superapp apps that use AI pay per request to the AI provider you connect, so before you set a price, do this calculation for your heaviest realistic user, not your average one:

1. Estimate how many AI requests a heavy user makes in a month.
2. Multiply by what each request costs you with your AI provider.
3. Subtract the store's commission (15% on the App Store for small developers in the Small Business Program) or your card processing fee (2.9% plus 30 cents per transaction with Stripe on the web) from your price.
4. What is left after AI cost and fees is your margin on that user. If it is negative, change the price, cap usage, or move work to a cheaper model.

On Apple devices, Apple's on-device Foundation Models can handle some tasks, such as summarizing or extracting text, without a per-request bill. Pricing structure also matters: RevenueCat found that 35% of apps now mix subscriptions with consumables or lifetime purchases, and that apps with cheap annual plans retain up to 36% of users after a year, versus 6.7% for high-priced monthly plans (RevenueCat). Usage-based credits are one common way to keep heavy AI users profitable.

App Store app or web app?

Superapp builds both, so the choice is about your users, not your tools.

What it actually takes, stage by stage

Is the app path right for you?

The people who make money from apps rarely do it with their first app. Levels had built and launched many products before Photo AI. Treat the first app as tuition, keep your costs low, and decide in advance how long you will give it.

How to choose an AI app idea that can make money

Superapp will build whatever you describe, so the decision that matters most happens before you open it. A money-making AI app idea passes all of these checks.

Which growth channels worked for real AI apps

Superapp's Selfdrive covers several of the channels in this table. The rest are worth knowing because each worked for at least one real app in this guide.

How to build and sell an AI app with Superapp

As a disclosure, Superapp and its growth agent, Selfdrive, are our products, so weigh this accordingly. Superapp builds native Swift apps for iPhone, iPad, Apple Watch, and Mac, and builds web apps, from a description. For App Store apps, it compiles on real Macs in the cloud, fixes its own build errors, prepares screenshots, metadata, and privacy declarations for submission, and lets you export the full Xcode project at any time. You can connect Superapp Cloud or your own Supabase project for data, use StoreKit or RevenueCat for subscriptions, and connect the AI provider of your choice for features like chat, transcription, or image generation. Plans start free with credits, then Pro at 25 USD a month, Business at 50 USD, and Max at 100 USD, which includes Selfdrive. Publishing on the App Store also needs Apple's 99 USD a year Developer account.

Selfdrive is the growth half. It runs Apple Ads, App Store optimization, creator video campaigns, and a website with SEO content for iOS apps, and emails a daily report with what it did and what needs your approval. It is in beta, it works with any App Store app you own, and it does not guarantee results, a point its own pages make clearly.

What Superapp does not do for you: pick a problem worth solving, decide your price, or make people want the app. Other routes exist too. Lovable and Replit build web apps, Flutter and React Native build for both app stores, and Xcode is free if you can code. The choice of tool matters much less than the choice of problem.

A practical first 30 days:

1. Week 1: Choose one problem. Check that people search for it in the App Store and Google. Write one sentence describing who the app is for and what it does for them.
2. Week 2: Build the first version in Superapp. One core screen that does the job, a paywall, and nothing else. Test it on your own phone daily.
3. Week 3: Put it in front of 10 people who have the problem. Watch where they get stuck. Fix the biggest issue. Submit to the App Store, or publish the web app.
4. Week 4: Start growth. Write a listing around the words people search, start a small Apple Ads test, and post short videos of the app doing its job. Track who pays and who comes back.

2. Sell AI Automation Services to Businesses

Many app builders start here, because automation work teaches you which problems businesses will pay to solve, and those can become apps later. The idea: small businesses know AI could save them time but do not know how to set it up. You build the automations for them, such as answering common customer emails, qualifying leads, summarizing calls, or generating weekly reports, and charge a setup fee plus a monthly retainer.

Demand is real. Upwork reported a 53% year-over-year increase in gross services volume from AI-related work in the third quarter of 2025 (Seeking Alpha on Upwork's earnings).

Why it is not trivial: this is a sales business first. You will spend more time finding clients and explaining value than building automations. Every client's systems are different, automations break when tools change, and the client will call you when they do. Price retainers to cover that maintenance.

Best for: people who are comfortable selling, and who know one industry well enough to speak its language.

Real AI service businesses, in their own words

Superapp is a product path, and services are the opposite path: you sell your time and skill first. These two stories show both the upside and the grind.

Aryan Mahajan, Zoro: from 180 USD chatbots to over 100,000 USD a month. Aryan describes starting with nothing: "I was in Dallas and out of ideas." He sent 47 proposals on Upwork in one day, each with a custom video showing a working prototype for that buyer, and his first client paid 180 USD for a real estate chatbot (Indie Hackers). His company now builds AI systems that run costly operational work for businesses, sold as a build fee plus a subscription, and the article reports over 100,000 USD a month in recurring revenue. Two lines capture what he learned: "Nobody with a real budget wants a chatbot," and "Show the machine, do not describe it." The road included maxed-out credit cards and, by his own account, saying yes to too much low-paid adjacent work early on.

Steve Jonson's AI voice agency model: 300 to 800 USD a month per client. Steve outlines an agency that sets up AI phone answering for small businesses, charging 800 to 2,000 USD for setup plus a 300 to 800 USD monthly retainer, with his estimate of hard costs at 50 to 100 USD a month per client (Indie Hackers). His pitch to the client is about lost revenue: "The math is stupid simple: these businesses lose $500-2K per missed call." His advice to builders: "Stop learning. Start shipping." For balance, the post models revenue targets rather than reporting a client count, and a commenter who runs a similar business describes margins tightening at around 12 clients because of billing and tool overhead.

Your first week: Pick one industry you can reach, such as dentists, real estate agents, or local gyms. List three tasks they repeat every day. Build one working demo of one of those tasks and show it to five businesses. Charge for the first project even if the price is small, because a paying client teaches you more than a free one.

Not for you if: You dislike sales calls, or you want income that does not depend on clients. Every automation business is a relationship business.

3. Freelance With AI as a Productivity Tool

Freelancing is also how many app builders, including people building with Superapp, pay the bills while an app grows. Freelancing with AI means doing the work you already do, such as writing, design, coding, translation, research, or video editing, faster and better with AI tools, so you can take on more clients or higher-value projects.

This is the fastest way to earn with AI, because the demand already exists and you only need to be good at the work. The same Upwork data on AI-related work shows clients are actively hiring for AI-assisted skills.

Why it is not trivial: if AI makes a task faster for you, it makes it faster for everyone, and clients know it. Prices for simple tasks like basic copywriting or translation fall as AI gets better. The freelancers who keep earning sell judgment and results, not output. Charge by the project or the outcome, not by the hour, or AI efficiency will cut your pay instead of raising it.

Best for: anyone with an existing skill who needs income soon.

What the data and freelancers say

Superapp users who freelance on the side should know what the research shows. Economists Xiang Hui and Oren Reshef of Washington University in St. Louis and Luofeng Zhou of New York University studied Upwork after ChatGPT's release and found that writing jobs fell 2% and monthly earnings for those freelancers fell 5.2% in the months that followed, describing "a causal relationship" with ChatGPT (TheWrap). Demand moved; it did not disappear: the same period saw Upwork's AI-related work grow quickly.

Maya Say, a freelance writer, described the shock directly: "I've been a freelance writer for 10 years." Her income "reached six figures last year before plummeting to almost zero two months ago" (Medium). Then she changed what she offered: "Well, I simply adjusted a few things, and my income started growing again, plus I'm doing more of the work I love."

Your first week: Write down the three kinds of work you already do well. For each, try doing it with AI tools and note where it saves real time. Rewrite your service description around the result, such as "first drafts in 24 hours," not the hours you spend. Raise your rate for new clients, since you now deliver faster.

Not for you if: Your skill is one AI can already do almost as well for most clients, with no judgment involved. In that case, move toward the work that needs judgment, or toward services and apps.

4. AI Consulting and Implementation in an Industry You Know

Consulting is also a good source of app ideas, since clients show you which problems they will pay to solve. Consulting is the expert version of automation services: you help a company decide where AI fits in its operations, choose tools, train staff, and roll it out. It pays best when you combine AI skills with real experience in one industry, such as law, healthcare administration, real estate, or logistics.

Why it is not trivial: credibility is everything. A company will not hand its processes to someone who learned AI last month and does not know its industry. Expect a long sales cycle, and expect to be judged on business results, not on how impressive the tools are.

Best for: professionals with years in one field who want to turn that experience into higher-paid advisory work.

Your first week: List the AI questions people in your industry ask you. Write a one-page plan for how AI could help with each, with tools, risks, and expected time savings. Offer it to one company you already know as a paid assessment.

Not for you if: You do not have experience in a specific field. Generic AI consulting is crowded, and companies hire people who already understand their business.

5. Create Content With AI Help

Superapp's growth agent, Selfdrive, uses short creator videos for a reason: content reaches people at scale. Making money from content with AI means using AI to research, script, edit, and produce videos, newsletters, or posts faster, then earning from ads, sponsors, affiliate links, or your own products.

Why it is not trivial: platforms are actively pushing back on low-effort AI content. From July 15, 2025, YouTube updated its Partner Program to demonetize "inauthentic" content that is mass-produced or repetitive (Plagiarism Today). Faceless channels that churn out AI videos are exactly what that rule targets. Content also takes months to earn anything, and most channels never reach monetization at all.

What works: using AI to make your own content faster, with your own knowledge, face, voice, or opinions at the center. AI should be the editor and assistant, not the author.

Best for: people with a point of view and the patience to publish consistently for a year.

Even the biggest creators are uneasy. In October 2025, MrBeast publicly wondered how AI-generated videos could affect the "millions of creators currently making content for a living," adding: "Scary times" (TechCrunch).

Your first week: Choose one topic you know and one format, such as a weekly newsletter or two videos a week. Use AI to research and outline, but write the conclusions yourself. Publish, then check what people actually read or watch.

Not for you if: You want income within months. Content usually takes a year of consistent publishing to earn meaningfully, and many channels never do.

6. Teach AI Skills

Many people want to use AI for something specific and do not know where to start. Teaching AI skills means courses, workshops, coaching, or corporate training on how to use AI for a specific job, such as AI for accountants, AI for real estate agents, or AI for teachers.

Why it is not trivial: general "learn ChatGPT" courses are everywhere, many of them free. The paid market is specific: a clear audience, a clear outcome, and a teacher with credibility in that field. Course content also goes out of date quickly as tools change, so expect to update it often.

Best for: people who already teach, train, or create content in a professional niche.

Joshua Wöhle's Mindstone shows what the paid end of this market looks like. Mindstone teaches non-technical employees to use AI at work through a 10-hour program delivered over four weeks, grew through a free community of in-person AI meetups, and the article reports over 200,000 USD in revenue in a single month, with clients including Hyatt and Pearson (Indie Hackers). Wöhle's most surprising lesson: "The biggest thing I learned over the past five years building Mindstone is that most people don't like learning." He used to see the job differently: "I used to believe our job was to kindle a fire, to inspire a love of learning." Paid AI training succeeds when it delivers a clear outcome for a busy professional, not when it makes AI exciting.

Your first week: Run one free workshop for a specific audience, such as "AI for real estate agents in 45 minutes." Ask attendees what they would pay to learn next. Build the paid version around the answers, not around what is exciting to teach.

Not for you if: You dislike updating material. AI tools change monthly, and a course that is six months old can already be wrong.

7. Sell AI-Assisted Digital Products

Superapp is about apps, but some people start smaller: templates, printables, planners, stock images, prompt packs, or design assets, made faster with AI and sold on marketplaces or your own site.

Why it is not trivial: this is one of the most crowded ways to make money with AI, because it is the easiest to start. Marketplaces also have rules. Etsy, for example, says "Sellers must disclose within their listing description if an item is created with the use of AI" (TechCrunch). And because the US Copyright Office has concluded that purely AI-generated material is not copyrightable, others may be free to copy your purely AI-generated designs.

What works: products where your own selection, design, and expertise are the value, sold to a specific audience you can reach.

Best for: designers and organizers who want a small, low-cost side income.

Two honest experiments show how hard this path is without distribution.

Matt Busel listed 130 products on Gumroad in 13 days and made 0 USD. Using AI coding tools, he built 130 products, including 88 small apps, to test the claim that AI lets anyone start a one-person business with no money and no audience (DEV Community). The result was zero sales and zero downloads of his free products. "I ran that experiment already," he writes. His conclusion is the clearest sentence about AI-made products anywhere: "130 products with no distribution is the same as 0 products." And on why it fails: "The AI made the shelf infinite."

Lars Montan launched 4 digital products in 3 weeks and made 0 sales. He sold Notion templates, a business system PDF, and a prompt pack on Gumroad and Etsy, promoted through LinkedIn posts three times a week (Indie Hackers). "LinkedIn is a viewing audience, not a buying audience," he concluded, and on the format itself: "A PDF sits in a downloads folder." He now plans to build interactive tools rather than static files and to check demand first.

Your first week: Search the marketplace for your product idea and read the reviews of the top sellers. Note what buyers complain about. Make one product that fixes those complaints, and decide where you will promote it before you make it, since the experiments above show that distribution decides everything.

Not for you if: You expect to list products and let marketplaces send buyers. They rarely do for new sellers.

8. Write and Publish With AI Assistance

Writing and publishing with AI means using AI to help research, outline, and edit books, guides, or newsletters, then selling them or building a paid audience.

Why it is not trivial: publishing platforms are flooded with low-quality AI-written books, readers have learned to avoid them, and platforms are adding disclosure and quality rules. A book still needs a real audience and a reason to exist. AI makes writing faster; it does not make an unknown author findable.

What works: nonfiction in an area you know, where AI speeds up the drafting and you supply the experience and accuracy.

Best for: experts with knowledge worth packaging, and writers who want a faster workflow.

Your first week: Pick a subject where you have real expertise and an existing audience, even a small one. Outline a short guide that answers one specific question better than anything currently available. Use AI to speed up drafting and editing, and check every fact yourself.

Not for you if: You plan to publish many AI-written books quickly. Readers and platforms have learned to spot them, and they rarely sell.

9. Sell Prompts, Custom GPTs, and AI Templates

This space moves quickly: what is a valuable prompt today is often a built-in feature tomorrow. Selling prompts means packaging prompts, custom GPTs, or AI workflows for a specific task and selling them.

Why it is not trivial: the market is small, prices are low, and AI tools keep getting better at understanding plain requests, which reduces the need for clever prompts. Most prompt sellers make very little.

What works: bundling prompts into a complete workflow for a specific profession, which is really a small digital product (see number 7) or the seed of an app (see number 1).

Best for: an experiment, not a plan.

Your first week: Turn your best prompt workflow into a small free resource and see if anyone uses it. If they do, the paid version should be a tool or app that does the work for them, which is where a builder like Superapp comes in.

Not for you if: You need meaningful income. Treat prompt selling as a way to test demand, not as a business.

10. Review and Recommend AI Tools

Many AI tools run affiliate or referral programs, which is why so many sites review AI tools. Making money this way means building a website, newsletter, or channel that tests and recommends AI tools, and earning commission on sign-ups.

Why it is not trivial: it is a content business with all the slowness of number 5, plus strong competition from established review sites. Earning commission also creates a conflict of interest that readers notice. The sites that last disclose their affiliate links and give honest negative reviews.

Best for: people who already test tools for their work and enjoy writing about them.

Your first week: Pick one category of AI tools you use for real work. Test three tools in it on the same task and write an honest comparison, including what each does badly. Disclose any affiliate links clearly.

Not for you if: You are not willing to write negative reviews. Readers trust review sites that criticize, and they leave the ones that only promote.

Which Way Fits You?

Superapp is the right tool for one row of this table. Find the row that matches your situation and start there.

Many people combine two: freelance or consult for income now, and build an app on the side, using what clients ask for as the idea.

Decide in Five Questions

Superapp is the right first step for some readers and the wrong one for others. Answer these five questions honestly, and the path usually picks itself.

1. Do you need money within the next 30 days? If yes, start with freelancing with AI or selling AI automation services. An app will not pay that fast.
2. Do you know a specific problem well enough to name the user in five words? If yes, the app path is open to you. If no, freelancing or services will show you which problems people pay to solve.
3. Will you market every week, not just build? If no, avoid the app and content paths, since both depend on distribution.
4. Do you have years of expertise in one field? If yes, consulting or teaching AI in that field may pay best, and can later feed an app idea.
5. Can you give one path six months before judging it? If no, choose freelancing, where results are visible in weeks.

Many people end up combining paths: services or freelancing to pay the bills, and an app built with Superapp on the side, using what clients ask for as the idea.

Every Story in This Guide, Side by Side

Superapp readers asked for real numbers, so here they are in one place. Each figure is as reported by the founder or the publication linked in its section.

A 90-Day Plan to Make Your First Money With AI

Superapp fits into days 31 to 90 of this plan for readers taking the app path. The plan is built so you earn something early and learn what people pay for before betting on a product.

How Much Can You Actually Make With AI?

Superapp can show you how fast an app can be built; nobody can honestly promise what it will earn. These are the real numbers available, and what they mean.

Anyone who promises a specific income from AI without knowing your skills, time, and market is selling something.

How to Spot a "Make Money With AI" Scam

Superapp's advice on this is simple: real businesses do not promise income. The make-money-with-AI space attracts schemes that sell the dream of passive income for an upfront fee.

In September 2024, the US Federal Trade Commission announced Operation AI Comply, a crackdown on deceptive AI claims. It included cases against business opportunity schemes that promised AI-powered passive income: the FTC said Ascend Ecom "has defrauded consumers of at least $25 million," and that FBA Machine "cost consumers more than $15.9 million," while another scheme promised an "AI-powered Ecommerce Empire." FTC Chair Lina Khan said: "Using AI tools to trick, mislead, or defraud people is illegal" (FTC announcement, via VKTR).

Mistakes People Make Trying to Make Money With AI

These apply to every path on this list, including building an app with Superapp.

• Starting with the tool instead of the problem. "I want to build an AI app" is not an idea. "Busy nurses need a faster way to write shift handover notes" is.
• Assuming AI is the moat. Everyone has access to the same models. Your edge is your knowledge, your audience, or your execution.
• Building before checking demand. Search the App Store, Google, and communities for the problem first.
• Pricing below cost. If each user's AI usage costs money, a cheap unlimited plan can make your best users your biggest losses.
• Ignoring distribution until launch. Plan how the first 100 users will find you before you build.
• Trying five paths at once. Pick one, give it a fixed amount of time, and measure it.
• Paying for a course or "system" that promises income. Spend that money on the tools and a small test budget instead.
• Quitting after the first attempt. Most successful app makers and creators had several failures first.

Frequently Asked Questions

What is the best way to make money with AI?
Building and selling an AI app has the highest ceiling, because subscription income can grow without more of your hours, and AI apps earn about twice the median revenue per install. It is also the hardest path. If you need income soon, freelancing with AI or selling AI automation services pays faster.

Can you really make money with AI?
Yes, but not passively and not automatically. People earn with AI by building apps, selling automation services, freelancing faster, consulting, creating content, and selling digital products. All of them take real work, and most attempts at each earn little at first.

How do beginners make money with AI?
Beginners usually start by freelancing with AI in a skill they already have, or by selling simple AI automations to local businesses. Building an AI app with a builder like Superapp is open to beginners too, since it does not require coding, but plan for months of marketing and iteration.

How much money can you make with an AI app?
It varies enormously. RevenueCat found the top 5% of new apps earn 8,880 USD after two years while the bottom 25% earn 19 USD or less, and most apps never reach 1,000 USD a month within two years. Top AI apps like Cal AI earn tens of millions, but they are rare exceptions.

Do I need to know how to code to build an AI app?
No. AI builders like Superapp build native Apple apps and web apps from a description, and handle compiling and App Store preparation. You still need to choose the problem, test the app, set the price, and market it.

Is it better to build an App Store app or a web app?
An App Store app gives you Apple's search audience, built-in payments, and trust, at a 15% commission for small developers in the Small Business Program. A web app keeps more of each payment, about 2.9% plus 30 cents per card transaction with Stripe, but you must find every user yourself. Many founders choose based on where their users already spend time.

How much does it cost to start making money with an AI app?
Superapp starts free with credits, then 25 USD a month on Pro. Publishing on the App Store needs a 99 USD a year Apple Developer account. Add your AI provider's usage costs and a small budget for early marketing, such as Apple Ads tests.

What is the fastest way to make money with AI?
Freelancing with AI in a skill you already have, because the demand exists and you can start this week. Selling AI automation services to businesses is the next fastest, though it depends on your ability to find clients.

What is the laziest way to make money with AI?
There is no lazy way that reliably works. The low-effort options, like selling AI-generated prints or prompts, are crowded and earn little, and platforms like YouTube demonetize mass-produced AI content. Offers that promise lazy AI income are a warning sign.

Is making money with AI on YouTube still possible?
Yes, if AI helps you make better content rather than replacing you. Since July 15, 2025, YouTube demonetizes inauthentic, mass-produced, or repetitive content, which targets low-effort AI channels.

Can I sell AI-generated art?
Often yes, but check each platform's rules. Etsy requires sellers to disclose items created with AI. The US Copyright Office has concluded that purely AI-generated material is not copyrightable, so others may be able to copy it.

Are "make money with AI" courses worth it?
Be careful. The FTC has sued business opportunity schemes that promised AI-powered passive income, including one it said defrauded consumers of at least 25 million USD. Learn from free, verifiable sources and spend money on tools and real tests instead.

Why do most AI apps fail to make money?
Usually not because of the code. Most fail because nobody finds them, they are too similar to existing apps, users leave after the first week, or AI costs eat the margin. RevenueCat's own conclusion is that AI alone does not drive success; differentiation does.

How do I get users for my AI app?
Start with App Store optimization and a website, then test Apple Ads and short creator videos. Superapp's Selfdrive agent runs these channels for iOS apps with your approval, but no tool can guarantee users; the app still has to be something people want.

How long does it take to make money with an AI app?
The first version can be built in days. Reaching meaningful revenue usually takes many months of marketing and improvement, and RevenueCat's data shows most apps do not reach 1,000 USD a month within two years. Plan your time and budget for a long effort.

Do AI apps make more money than regular apps?
On average, per user, yes. RevenueCat found most AI apps earn above $0.63 per install after 60 days, about twice the overall median of $0.31. That does not mean most AI apps succeed; outcomes are extremely uneven.

What kind of AI app makes the most money?
Apps where AI does visible, repeated work for a specific person, such as logging calories from a photo or tracking expenses with less typing, tend to earn. A general chat app with a theme struggles, because it competes with free AI assistants.

How do I price an AI app?
Work out what your heaviest realistic user costs you in AI requests, then price above that after the store's commission or card fees. RevenueCat found that 35% of apps mix subscriptions with consumables or lifetime purchases, and usage credits are a common way to keep heavy AI users profitable.

Should I build one app or many?
Both work. Max Artemov built a 30-app portfolio around single App Store keywords and reached about 22,000 USD a month, while Florian Vates earns about 90% of his income from one app. More apps only help if each has a way to be found.

Is it too late to make money with AI apps?
No, but it is no longer enough to be first with an AI feature. The apps that earn now are differentiated by the problem they solve, the users they reach, and how well they are marketed.

Can I make money with AI without coding?
Yes. AI builders like Superapp build apps without code, and services, consulting, teaching, and content all use AI tools without programming.

How do AI automation agencies get clients?
Real examples include sending tailored proposals with a working prototype, as Aryan Mahajan did with 47 Upwork proposals in one day, and pitching the cost of a problem, such as missed calls. Sales skill matters more than technical skill.

Has AI reduced freelance income?
For some kinds of work, yes. A study by researchers at Washington University in St. Louis and New York University found monthly earnings for affected Upwork freelancers fell 5.2% after ChatGPT's release, while AI-related work on Upwork grew quickly.

Why do AI digital products often get zero sales?
Because AI makes them easy to produce, so marketplaces fill up and buyers cannot find yours. Matt Busel listed 130 AI-built products on Gumroad and made zero sales, concluding that "130 products with no distribution is the same as 0 products."

What is an AI wrapper, and why is it risky?
An AI wrapper is a product that adds a thin interface over another company's AI model. It is risky because competitors can copy it quickly and the model provider can add the same feature itself. Durable AI businesses own something extra, such as data, workflow, an audience, or expertise.

References

• RevenueCat: The State of Subscription Apps 2025

• TechRepublic: Most apps fail to hit 1,000 USD monthly revenue within two years

• Apple Ads: Advanced

• Apple Developer: App Store Small Business Program

• Apple: App Store Transparency Report 2025

• Apple: App Review Guidelines

• Stripe: Pricing

• Fortune: Cal AI founder Zach Yadegari

• Indie Hackers: Pieter Levels passes 100,000 USD a month with Photo AI

• TechCrunch: Base44 sells to Wix for 80 million USD

• Seeking Alpha: Upwork AI-related work growth

• Plagiarism Today: YouTube targets inauthentic content

• TechCrunch: Etsy's policy on AI-generated items

• Copyright Alliance: Summary of the US Copyright Office AI report, Part 2

• VKTR: FTC announces crackdown on deceptive AI claims and schemes

• Indie Hackers: Partnering with a content creator to hit 50k a month (MonAi)

• MRR Story: 6,772 USD MRR with 20 USD a day in Meta ads (RotateProduct)

• Indie Hackers: How I grew my app's monthly revenue from 200 to 3K USD

• Indie Hackers: From failed app to 30-app portfolio making 22K a month

• Indie Hackers: AI tool for agencies, 40 days, zero paying strangers

• Indie Hackers: Your AI product is not a real business

• Indie Hackers: From broke and idealess to a 100K-a-month AI services company

• Indie Hackers: Building a profitable AI voice agency

• TheWrap: Study finds ChatGPT drives down freelance work and pay

• Medium: ChatGPT halved my income as a freelance writer (Maya Say)

• TechCrunch: MrBeast says AI could threaten creators' livelihoods

• Indie Hackers: A 200K-a-month business teaching AI competency (Mindstone)

• DEV Community: 130 products on Gumroad, 0 sales (Matt Busel)

• Indie Hackers: 4 digital products in 3 weeks, 0 sales (Lars Montan)

• Superapp: Selfdrive

• Superapp: How to make money from an app

• Superapp: Simple app ideas that make money

• Superapp: How to build an app