How to build a simple business app without a developer

A custom business app used to mean hiring a developer, spending months on requirements, and paying thousands of pounds for something that didn't quite do what you needed. Or it meant wrestling with off-the-shelf software that was close but not right, paying for features you didn't need while missing the ones you did.

That calculation has changed. AI tools can now build functional, custom business applications from a plain text description. Not perfect, not always polished, but working. And working in an afternoon, not in three months.

What kind of app are we talking about?

Not a mobile app in the App Store. Not a SaaS product. A web-based tool that you and your team use to run part of your business. Things like:

A job tracker for a trades business. A plumber, electrician, or builder managing 10 to 30 jobs at any time needs to know the status of each one, what's been quoted, what's been invoiced, and what's overdue. A spreadsheet works up to a point. A proper tracker with columns, filters, and status fields works better.

A client portal for an accountant. A simple web page where clients can upload documents, see the status of their accounts, and access completed work. Not a full accounting platform. Just a clean, organised front end for client communication.

A booking system for a hairdresser. A form where clients can request an appointment, see available slots, and get a confirmation. Connected to your calendar. No phone tag, no missed enquiries.

These are real things that real businesses need, and they are well within what AI tools can build right now.

Which tool to use

Bolt (bolt.new): start here for simple tools

Bolt is the fastest path from idea to working app. If your tool is relatively straightforward, go to Bolt first. It builds quickly, the output is good, and you can iterate by typing what you want to change.

The limitation is complexity. Bolt handles simple tools well. If you need user accounts, multiple permission levels, or complex data relationships, you will hit its ceiling fairly quickly.

Google AI Studio with Antigravity: more power, still free

Google AI Studio's Antigravity feature is more capable than Bolt when you need application logic. User accounts, databases, integrations with other services. It takes a bit longer to get a result but produces something more substantial. And it is free, which removes the cost barrier for experimentation.

Good for: client portals, tools with user logins, anything that needs to store data across sessions.

Lovable (lovable.dev): the most capable no-code builder

Lovable is the most capable of the three for complex applications. It handles multi-user apps, custom data models, and integrations well. It is also more polished in terms of the output. The trade-off is that it takes longer to get started and has a steeper learning curve.

Good for: anything that needs to feel professional, handle multiple users, or integrate with external systems.

How to do it

Step 1: Write a proper brief

The quality of what you get out is directly related to the quality of what you put in. Spend 10 minutes writing this before you open any tool.

Describe:

• What the app does (one sentence)
• Who uses it and how often
• What data it needs to store or show
• What actions users need to take
• What it should look like (simple, professional, branded)

Example for a job tracker:

Build a job tracking tool for a plumbing business. The owner needs to see all active jobs in a table with columns for: job name, client name, address, status (quoted, booked, in progress, invoiced, paid), and notes. The owner should be able to add new jobs, update the status by clicking on it, and filter by status. Keep it simple and clean. Mobile-friendly.

That is a real brief. It will get a real result.

Step 2: Build it

Paste your brief into whichever tool you have chosen. Watch it build. The first version will probably be 70 to 80 percent of what you need. That is fine.

Step 3: Iterate

Tell it what to change. "Make the status column a dropdown", "add a search box", "change the colour scheme to match our brand (dark blue and white)", "add a field for the date the job was booked".

This part takes time. Plan for 1 to 2 hours of back and forth for something useful.

Step 4: Test it properly

Use it as you would in real life. Add real data. Try the edge cases. What happens if a field is left blank? What happens if you add 50 jobs? Find the problems now, before you rely on it.

Step 5: Share it or deploy it

Most of these tools let you share a link directly or deploy to a hosted URL. For internal tools, sharing a link is usually enough. For something client-facing, you may want a custom domain.

Honest limitations

AI-built tools are not the same as professionally built software. They can be fragile. They may break when you try to do something unexpected. They lack the testing and error handling that a good developer would build in.

For low-stakes internal tools, this is usually fine. For anything handling sensitive client data, payments, or business-critical processes, be cautious. A working proof of concept is valuable. Relying on it for something that cannot fail is a different matter.

The other limitation is that iterating through AI can take longer than you expect. The tool does not always understand exactly what you mean, and you may spend time correcting things that have gone slightly wrong. That is still faster and cheaper than hiring a developer for the same output.

Start simple. Build the thing that solves the most pressing problem first. You can add to it later.