Can You Patent an Algorithm and What Does It Actually Take to Qualify
Algorithms are generally understood and broadly defined by Merriam-Webster’s dictionary as a “step-by-step procedures for solving problems or accomplishing some ends.” In…
[INDUSTRY]AI and machine learning
AI and machine learning patent attorneys for tech startups
We protect the model, the method behind it and the product name, claimed as a technical improvement. One flat fee for each step.
Your attorneys
Andrew RapackeManaging Partner · Registered Patent AttorneyFormer Naval Engineering Officer. M.S. in electrical and electronics engineering.
Chris WattRegistered Patent Attorney · Former USPTO ExaminerSoftware, AI and machine learning, and MedTech. A full-stack software engineer.
[01]Who it's for
For the people who build with AI.
From a two-person model team to a company adding AI to a product it already sells, the work starts the same way. Find out what is new, then protect it before it ships or gets published.
AI startups
Founders whose product is the model or the pipeline around it, protected before the first investor meeting or launch.
Companies adding AI features
Teams building AI into a software product they already sell, protecting the features that set it apart.
Research and engineering teams
Engineers with new training methods, architectures or ways to run models faster on less hardware.
Applied AI teams
Models that read images, signals or transactions, with claims built around the result they produce.
[02]What's protectable
Four things worth protecting in an AI product.
Most AI products carry more than one kind of protection. A patent on how the system works, a design patent on how its interface looks and a trademark on its name each stop a different kind of copy.
01Utility patent
The model and how it is trained
New architectures, training methods and ways of preparing data, claimed as the technical steps that make the model work better rather than the math alone.
- Model architectures
- Training and fine-tuning methods
- Data preparation pipelines
- Running models on less compute
Flat fee, quoted before work begins02Utility patent
AI applied to a real problem
A model put to work on a specific task, such as reading a scan, flagging fraud or controlling a machine, claimed around the technical result it produces.
- Computer vision
- Natural language processing
- Prediction and anomaly detection
- AI agents and automation
Flat fee, quoted before work begins03Design patent
The interface people see
The ornamental design of an app screen, an icon or a chat interface, protected for 15 years from grant.
- App and dashboard screens
- Icons
- Conversational interfaces
USPTO denies it, full refund*04Trademark
The product and model name
Clear and file the name before it goes on your website, your app store listing and your API docs, so the name you build a user base on is one you can keep.
- Product and model names
- Company names
- Logos
Approved, or your money back*
Some things stay secret
Training data, model weights and tuning settings no one can learn from your product may be worth more as trade secrets than as published patents. Your attorney tells you which on your first call.
[03]The flat-fee path
Search, provisional, utility. Flat fees at every step.
Every step is quoted before any work begins, and the search fee is credited toward a design or non-provisional application. Most AI founders file a provisional before a launch, a demo or a paper, then the full application within 12 months.
01Start here
Patentability search
$999
Patentable or your money back. Credited toward a design or non-provisional application
- An invention discovery call with your attorney
- A search of US and foreign patents and published applications
- A written patentability report and a review call
Report in about 5 days
Not patentable, full refund*02Then
Provisional patent
$2.5k–3.5k
- A strategy call to capture how the system works
- Flowcharts and system diagrams included
- You review and approve it before filing
Patent pending the day it is filed
Patentability issues, full refund*03Within 12 months
Non-provisional patent
$7k–12k
Drawings and USPTO filing fees included
- Method and system claims written for examination
- Formal drawings and USPTO filing fees
- Complex systems quoted flat on your call
Filed before your 12-month deadline
Flat fee, not covered by the guarantee
Alongside the patent
- Design patentFor an app screen or interface, office actions included$3,200USPTO denies it, full refund*
- TrademarkFor the product name, USPTO fee for one class included$2,000Approved, or your money back*
*Refunds subject to the RLG Guarantee Policy in your engagement letter. Utility (non-provisional) patents are quoted flat and are not covered by the guarantee. How The RLG Guarantee works
[04]AI and the USPTO
Patent rules for AI. Plan for them from the start.
A patent decides who can make and sell the technology. For AI inventions the USPTO also asks two questions worth planning for, whether the claims are more than an abstract idea and whether a person, not the model, is the inventor.
PatentWho can make and sell it
- $99901$999SearchPatentable, or your money back
- File first02File firstProvisionalPatent pending, before any disclosure
- Month 1203Month 12Full applicationWithin 12 months
- USPTO04USPTOExaminationThe USPTO checks eligibility and novelty
- 05GrantedRights you can enforce
ProductWhen others see it
- Use an NDA01Use an NDAPrototypeModels, data and test runs
- Disclosure risk02Disclosure riskPaper or preprintResearch made public
- Disclosure risk03Disclosure riskBeta or demoUsers see it working
- 04LaunchApp, API or platform
- 05Next versionNew features, new filings
File before you publish or demo
A preprint, a conference talk, a public code repository or a live demo can make an invention public. The US gives you one year after your own disclosure, but most other countries do not. File at least a provisional first.
An algorithm alone is not enough
Under 35 U.S.C. 101 and the Alice decision, math or a model described on its own is an abstract idea. Claims tied to a technical improvement, such as faster training or a more accurate reading, are what the USPTO looks for.
Can you patent an algorithmPeople invent, AI is a tool
Only natural persons can be named as inventors. The USPTO's November 2025 guidance treats AI like any other tool, so keep records of who conceived each part of what you claim.
Can AI be an inventorKnow where your data came from
Training data, open source models and the licenses on both affect what you own and what you can file on. Check them before you file, raise money or sign a partnership.
AI training data and IP
General information, not legal advice. Eligibility rules for AI change through court decisions and USPTO guidance, so your attorney checks the current standard against your claims.
[05]Related services
Services for AI and machine learning companies.
One firm for the patent, the brand and the agreements around them.
- AI patentsPatent the model, the training and the AI product
- Software patentsProtect the software and methods around your model
- Provisional patents, covered by The RLG GuaranteePatent pending while you train and test
- Design patents, covered by The RLG GuaranteeProtect how the interface looks
- Trademarks, covered by The RLG GuaranteeRegister the product and company name
- Patent office actionsAnswer eligibility and prior art rejections
- IP licensing agreementsLicense a model or technology in or out
- Privacy and data protectionPolicies for the data your models use
[06]Questions
Before you file.
Still deciding? Your first call is free and comes with a straight answer on whether to file.
Book A Free Strategy CallCan you patent AI?
Yes, when the claims cover a specific technical improvement rather than an abstract idea. A new way to train a model, run it on less hardware or apply it to a real task can qualify. A model described only as math, or a business process with AI added to it, usually does not. The patentability search tells you where yours stands before you file.
Can an AI system be named as an inventor?
No. US law requires inventors to be natural persons, as the Federal Circuit held in Thaler v. Vidal in 2022, and the USPTO treats AI as a tool the inventor uses. Inventions made with AI can still be patented when a person conceived what is claimed, so keep a record of who did what as you build.
Should I patent my model or keep it a trade secret?
It depends on what a competitor can see. Features users can see or work out from the product are often better patented. Training data, model weights and tuning that never leave your servers can stay trade secrets for as long as they stay secret. Many AI companies use both, and your attorney weighs it with you on the first call.
What does a machine learning patent attorney do?
Finds what is new in your system with a patentability search, drafts claims around the technical improvement, answers USPTO rejections and builds a portfolio that grows with the product. When someone copies a patented feature, the same firm handles enforcement through patent litigation.
What other legal work does an AI startup need?
Usually business formation, investor agreements, employee and contractor agreements that assign IP to the company, customer and vendor contracts, and NDAs for partners. AI companies also need privacy policies and data agreements that fit laws such as the GDPR and CCPA. We handle these on flat fees alongside the patent work.
How much does an AI patent cost with your firm?
Flat fees at every step. A $999 patentability search, a provisional application at $2,500 to $3,500 depending on the invention, and a non-provisional application at $7,000 to $12,000 with drawings and USPTO filing fees included. Every fee is quoted before work begins, and the search fee is credited toward a design or non-provisional application.
What does The RLG Guarantee cover for AI and machine learning?
The patentability search, the provisional application, design patents and trademarks. If the search finds your invention is not novel, you get a full refund or another search. If patentability issues come up on your provisional, a full refund or more searches. If the USPTO denies your design application, you get a full refund. If it refuses your trademark, you get a refund or a new mark. Non-provisional applications are not covered. Refunds are subject to the RLG Guarantee Policy in your engagement letter.
[07]Learning Center
Go deeper on AI patents.
Patent Eligibility for AI Inventions: What Inventors Need to Know
Learn the rules for patent eligibility for AI inventions, current USPTO guidance through 2026, and how to draft claims that survive § 101. Book a free IP strategy call.
Choosing An AI Patent Attorney: Everything You Need To Know
As an AI Patent Attorney, it’s clear that artificial intelligence (AI) is permeating almost every sector, from ChatGPT to smart indoor farming to speech recognition…
Can AI Be an Inventor? Intellectual Property in the AI Era
Can AI be an inventor? As everyone around the world gets excited about the new age of artificial intelligence (AI), innovation around AI has exploded. According to IBM…
Using Trade Secret Protection for AI IP
Artificial intelligence innovations often consist of several elements that would be best protected by trade secret. A trade secret is any business information that…
The Importance of Intellectual Property for AI Makers
AI makers, whether building foundational models, AI agents, or AI wrappers, need patents, copyrights, trade secrets, and trademarks working together. Here is how each protects artificial intelligence, and where U.S. law stands on AI inventors and authors.
[08]Client reviews
Founders who filed with us.
4.7 across 614 public reviews on Google, Trustpilot and Avvo.
#1Top reviewed IP law firm*
Read client reviews“A detailed analysis and report was provided once the search was completed. They answered all of my questions in a timely manner and advised me on the best strategies moving forward.”
Brittany LongPrior art search · Client review “They traversed the legal maze and minefield.”
Peter C.Patent · Trustpilot “I always felt confident that my intellectual property was in good hands.”
JakePatent · Client review
Reviews reflect individual client experiences. Quotes are excerpts. Prior results do not guarantee a similar outcome. *#1 top reviewed means the most public client reviews on Google, Trustpilot and Avvo of any US patent or trademark firm.
Protect the model before you publish it.
Talk through your AI with a registered patent attorney. The call is free, with no obligation.