What You Can Build Next — And What Should Wait

"Help me choose the right next project category.
I have already completed a tiny first win.
Now I want to map what kind of project makes sense next.
My idea is: [describe it]
Place this idea in one of three buckets:
1. realistic now

Start Here told you to pick a tiny, boring, useful first win. That advice still stands.

This lesson is the wider map.

The interesting question is no longer just "Can AI build real software?" The answer is yes. The more useful question is:

What kinds of products make sense now, what kinds make sense next, and what kinds should still wait?

That distinction matters because the category of project shapes the category of risk.

Landing Pages and Marketing Sites

This is the easiest starting point and where many vibe coders begin.

What you can build: A responsive website for your business, side project, or personal brand, complete with contact forms, email signup, and mobile-friendly design.

Why it matters: Before AI coding tools, you either paid a web designer, used a cookie-cutter template, or spent weeks learning HTML and CSS. Now you can describe your vision and have a custom site in an afternoon.

Try this prompt with Bolt or Lovable:

Build a landing page for a dog walking business called "Happy Tails." Include a hero section with a headline and call-to-action button, a section showing three service tiers (basic, premium, and VIP), customer testimonials, and a contact form. Use warm, friendly colors and include space for dog photos.

You'll have a working, professional-looking site in about two minutes.

Internal Business Tools

This is where vibe coding gets genuinely transformative.

What you can build: Custom dashboards, inventory trackers, appointment schedulers, invoice generators, employee directories, project management boards, and any other internal tool your business needs.

Why it matters: Most businesses run on a patchwork of spreadsheets, sticky notes, and tools that almost do what they need. Now the person who understands the problem best can often build the missing solution.

Real-world example: A small accounting firm needed a tool to track which clients had submitted their quarterly documents. Off-the-shelf software was either too simple or too expensive. Using Cursor, the firm's office manager built a custom tracking dashboard in a weekend. It shows which clients are current, which are overdue, and sends automatic email reminders.

SaaS Applications

Yes, people are building software-as-a-service products — the kind that charge monthly subscriptions — using vibe coding.

What you can build: Web applications with user accounts, subscription billing, dashboards, data storage, and all the features you'd expect from a "real" SaaS product.

Why it matters: The barrier to testing a SaaS idea used to be massive. You needed to hire developers, spend months building, and invest significantly before knowing if anyone would pay. Now you can build a working prototype in days and put it in front of potential customers.

Typical SaaS features vibe coders are building:

• User registration and login
• Subscription payments with Stripe
• Dashboard with charts and data visualization
• Admin panels for managing users and content
• Email notifications
• API integrations with other services

Mobile Apps

Building mobile apps traditionally required learning Swift (for iPhone) or Kotlin (for Android), or using cross-platform frameworks like React Native that still demanded programming knowledge. Now you can describe your app and generate it.

What you can build: Cross-platform mobile apps that work on both iPhone and Android, including features like push notifications, camera access, GPS, and offline storage.

Why it matters: The app stores are still a massive distribution channel. If your idea is best served as a mobile experience, vibe coding can get you there.

Caveat: Mobile apps are still trickier than web apps with AI tools. The development-to-deployment pipeline is more complex (app store submissions, device testing, etc.). Start with a web app and consider mobile once your idea is validated.

Browser Extensions

One of the most underrated things you can build.

What you can build: Extensions for Chrome or Firefox that modify websites, add features to tools you use daily, automate repetitive browser tasks, or collect data.

Why it matters: Browser extensions can solve very specific personal or professional pain points. And they're one of the simpler things to build with AI because they're relatively small and self-contained.

Try this prompt with Cursor or Claude:

Build a Chrome extension that adds a "Copy as Markdown" button to any webpage. When clicked, it should convert the selected text on the page to Markdown format and copy it to the clipboard.

Automations and Integrations

What you can build: Scripts and workflows that connect different tools, automate repetitive tasks, process data, generate reports, and generally make your digital life more efficient.

Why it matters: If you find yourself doing the same thing on a computer more than three times, it can probably be automated. Vibe coding makes those automations accessible to everyone.

Examples:

• A script that pulls data from your CRM every morning and sends a summary email
• An automation that monitors a website for changes and sends you a text
• A tool that reformats CSV files from one vendor's format to another's
• A bot that posts to multiple social media platforms simultaneously

Data Dashboards and Visualizations

What you can build: Interactive dashboards that pull data from spreadsheets, databases, or APIs and display it as charts, graphs, and tables.

Why it matters: Decision-makers need data presented clearly. Building custom dashboards used to require hiring a developer or buying expensive BI tools. Now you can build exactly the dashboard you need.

What's NOT Realistic (Yet)

Honesty is important. Here's what's still hard to build purely through vibe coding:

• High-security financial systems — Banking, trading platforms, anything where a bug means losing money
• Real-time multiplayer games — The complexity of game physics and networking is still very challenging
• Large-scale enterprise software — Systems used by thousands of employees with complex business logic
• Performance-critical applications — Things that need to process massive amounts of data very quickly

These aren't impossible, but they require deep technical knowledge that goes beyond describing what you want. The AI can help, but you'd need significant programming experience to guide it.

The Pattern

Notice the pattern in everything above? The best vibe coding projects share these traits:

1. The builder understands the problem deeply — You know exactly what the tool should do because you've lived with the problem
2. The scope is focused — It does one thing well rather than trying to do everything
3. It solves a real need — Not a solution looking for a problem, but a response to genuine pain

You don't need to be a programmer to have these qualities. In fact, sometimes not being a programmer is an advantage — you're focused on the problem, not the technology.

Try this now

• Take one idea you care about and place it in one of three buckets: now, next, or later.
• Decide whether the best near-term shape is a landing page, internal tool, automation, browser extension, or simple web app.
• Write down what risk makes the idea a bad fit for right now if it belongs in later.

Prompt to give your agent

"Help me choose the right next project category. I have already completed a tiny first win. Now I want to map what kind of project makes sense next.

My idea is: [describe it]

Place this idea in one of three buckets:

1. realistic now
2. realistic next after more practice
3. not realistic yet

Then explain the safest product shape for the current stage and the main reason not to jump ahead."

What you must review yourself

• Whether you genuinely understand the problem you are trying to solve
• Whether the project category matches your current operational skill, not just your ambition
• Whether the next move should be a better version of a small app instead of a bigger class of app
• Whether you are building something useful rather than something merely impressive-looking

Common Mistakes to Avoid

• Assuming "possible" means "appropriate right now." Category fit matters as much as technical possibility.
• Jumping into high-risk categories too early. Security-heavy and scale-heavy projects still punish beginners.
• Using a successful tiny win as proof you should now build a platform. Confidence should expand judgment, not erase it.
• Mistaking "AI can generate it" for "I can support it." Ownership still matters after generation.

Key takeaways

• Project categories come with different risk levels, operating demands, and support costs
• Web apps, internal tools, automations, and extensions are still the strongest early bets
• Deep problem understanding often matters more than deep technical knowledge
• The right question is not just "can AI build this?" but "should I build this now?"

What's Next

We've painted an exciting picture. Now it's time for some real talk. In the next lesson, we'll have an honest conversation about what AI can and cannot do for you — so you go in with realistic expectations and avoid the most common pitfalls.