Is My Idea Patentable? A Simple Framework

You’ve got what feels like a million-dollar idea. Before you hire a patent attorney, there’s a fundamental question: Can this actually be patented?

Not everything can. The patent office rejects more applications than it approves. Understanding the requirements upfront saves you thousands in wasted fees.

Here’s a simple framework to evaluate your idea.

The 4 Requirements for a Patent

To be patentable, your invention must be:

Novel (new)
Non-obvious (not a trivial variation)
Useful (has practical application)
Patentable subject matter (the right type of thing)

Let’s break down each one.

Requirement 1: Novelty (Is It New?)

Your invention must be different from everything that came before.

What counts as “prior art”:

Existing patents
Published patent applications
Academic papers
Products on the market
Public demonstrations
YouTube videos
Your own public disclosure (careful!)

The test: If someone, somewhere, has already disclosed your exact invention — you can’t patent it.

Important nuance: “Novel” doesn’t mean completely unprecedented. If you combine known elements in a new way, or apply existing technology to a new problem, that can be novel.

Example:

NOT novel: “A chair with four legs” (exists)
Potentially novel: “A chair with four legs that automatically adjusts height based on occupant weight using pressure sensors” (if not already done)

Requirement 2: Non-Obviousness (Is It Inventive?)

This is where most ideas fail.

Even if your invention is technically “new,” it might be obvious. The test: would a person skilled in your field easily come up with this solution?

Signs your idea might be obvious:

It’s a simple combination of existing solutions
Industry publications have suggested something similar
It’s the “natural next step” in the technology
Competitors could easily think of it

Signs your idea might be non-obvious:

Experts tried and failed to solve this problem
It produces unexpected results
It goes against conventional wisdom
It solves a problem people didn’t think was solvable

Example:

Obvious: “Put a phone mount on a tripod” (trivial combination)
Non-obvious: “A phone mount with AI-powered subject tracking that anticipates movement” (unexpected technical solution)

Requirement 3: Usefulness (Does It Work?)

Your invention must have a specific, practical use. This is the easiest requirement to meet.

What qualifies:

It performs a function
It solves a problem
It has industrial or practical application

What doesn’t qualify:

Purely theoretical concepts with no implementation
Things that violate laws of physics
Aesthetic designs (those get design patents, not utility patents)

You DON’T need:

A working prototype
Commercial viability
Market demand

You just need to show it could work and would be useful if it did.

Requirement 4: Patentable Subject Matter (Is It the Right Type?)

Not everything is patent-eligible, even if it’s new, non-obvious, and useful.

What CAN be patented:

Machines (devices, apparatus)
Processes (methods, steps)
Manufactures (products)
Compositions of matter (chemicals, materials)
Improvements to any of the above

What CANNOT be patented:

Laws of nature (gravity, E=mc²)
Abstract ideas (mathematical formulas alone)
Natural phenomena (a newly discovered plant species)
Purely mental processes (thinking, without implementation)

The gray area: Software and business methods

Software patents are tricky. Pure algorithms aren’t patentable, but software that produces a “technical effect” or improves computer functioning often is.

Usually patentable:

Software that controls a physical device
Algorithms that improve computer performance
Technical solutions to technical problems

Usually NOT patentable:

Generic business methods implemented on a computer
“Do [old thing] but on the internet”
Pure data manipulation without technical effect

Quick Self-Assessment

Answer these questions honestly:

Novelty:

Have you searched Google Patents for similar inventions?
Have you found anything that does exactly what yours does?
Did you publicly disclose your idea before filing? (Social media, pitch events, etc.)

Non-obviousness:

Would an engineer in your field say “why didn’t I think of that?”
Does your solution produce unexpected results?
Have others tried and failed to solve this problem?

Usefulness:

Does it do something practical?
Can you explain what problem it solves?

Subject matter:

Is it a machine, process, manufacture, or composition?
If software: does it have a technical implementation, not just a business process?

Red Flags (Your Idea Might Not Be Patentable)

🚩 “It’s obvious in hindsight” If everyone says “that’s obvious!” after hearing it, it might actually be obvious.

🚩 “I saw something similar in a research paper” Academic publications are prior art. If a paper describes your concept, even if no product exists, it may block your patent.

🚩 “I posted about it on social media” Your own public disclosure can be prior art. In the US, you have a 1-year grace period, but internationally you don’t.

🚩 “It’s just software that does X” Pure software ideas face extra scrutiny. The more “technical” and less “business method,” the better.

🚩 “Lots of companies do something similar” Existing products suggest prior art. Your specific implementation might still be patentable, but you need to be careful about claim scope.

What You Can Do If Your Idea Seems Patentable

Step 1: Document everything Write down your invention in detail. Include:

What it is
How it works
What problem it solves
What’s different from existing solutions
Date you conceived of it

Step 2: Do a thorough search Use Google Patents, USPTO databases, and ideally an AI patent search tool to find prior art. Don’t skip this.

Step 3: Evaluate your search results If you find close prior art, you’re not done — but you need to understand exactly what’s already claimed.

Step 4: Consider a professional opinion A patentability search from a patent attorney or AI tool gives you a more rigorous assessment before you commit to filing fees.

What If Your Idea ISN’T Patentable?

Not the end of the world. You have options:

1. Modify the invention Can you add something that makes it non-obvious? A new feature, a different implementation?

2. Protect it differently

Trade secrets (keep it confidential)
Copyrights (for code, designs)
Trademarks (for branding)
First-mover advantage (speed to market)

3. License instead of compete If someone else has the patent, maybe you can license it and build your business anyway.

4. Move on Not every idea is worth patenting. Sometimes the right move is to focus your energy on something more protectable.

Common Questions

“Can I patent an improvement to someone else’s invention?” Yes. Improvement patents are common. But you might still need a license from the original patent holder to practice your improvement.

“Can I patent something I found in another country?” Only if it wasn’t publicly disclosed. Foreign patents count as prior art.

“What if my idea is really simple?” Simplicity isn’t disqualifying. The paper clip is patented. But simple + obvious = not patentable.

“How much do I need to describe?” Enough that someone in your field could build it. Vague concepts aren’t patentable — specific implementations are.

Bottom Line

Most ideas that feel patentable at first glance face challenges when you dig deeper. The key requirements — novel, non-obvious, useful, and proper subject matter — seem simple but have nuance.

Before spending $10,000+ on a patent application:

Search for prior art thoroughly
Honestly evaluate non-obviousness
Get a professional patentability assessment

Better to know now than after you’ve spent the money.

Curious about your idea’s patentability? Get an AI assessment →