The Protection Playbook for AI and Data Startups

For AI and data-driven startups, intellectual property is often one of the company’s most valuable assets. Yet many startup founders struggle to determine what should be protected, how it should be protected, and whether patents are always the right answer.

During a recent Swiss Startup Association webinar, Kamran Houshang Pour, IP Expert and Senior IP Trainer at the Swiss Federal Institute of Intellectual Property (IPI), and Thomas Kuster, Attorney at Law and Notary at LEXR, shared practical insights into the IP strategies that help AI startups build and defend their competitive advantage.

Start with your competitive advantage

Every startup is different. Some companies create value through proprietary algorithms, others through unique datasets, technical know-how, source code, workflows, or brand recognition. As a result, there is no universal protection strategy. Instead of asking “How do I patent my startup?”, founders should first ask:

• What makes our company difficult to copy?
• Which assets create our competitive advantage?
• Can competitors reverse engineer what we are doing?
• Would disclosure strengthen or weaken our position?

The answers to these questions determine whether patents, trade secrets, contracts, copyrights, trademarks, or a combination of these tools make the most sense.

Patents: Powerful, but not always the answer

Patents remain one of the strongest forms of intellectual property protection. They provide a temporary monopoly over a technical invention, allowing companies to prevent others from using the same technology. To qualify for patent protection, an invention must be:

• New worldwide
• Inventive and non-obvious
• Technically feasible
• Technical in nature

One of the most important lessons for startup founders is that novelty can be lost surprisingly easily. Publishing details of an invention before filing a patent application can destroy patentability, even if the publication comes from the founder themselves.

The speakers also addressed a common misconception around software. While software “as such” is generally not patentable in Europe, software that creates a technical effect can be. For example, accounting software that performs calculations is unlikely to qualify. However, software that controls a technical process, such as a braking system in a vehicle, may be patentable because it solves a technical problem.

Why startups pursue patents

Patents are not only about preventing competitors from copying technology. For many startups, patents also serve important strategic purposes:

• Demonstrating innovation to investors
• Increasing company valuation
• Supporting future licensing opportunities
• Creating barriers to market entry
• Strengthening acquisition potential

Particularly in deep-tech sectors, investors often view patents as an important signal that a company has developed defensible technology. However, patents also come with costs, filing requirements, enforcement obligations, and public disclosure. Startup founders should therefore carefully evaluate whether patent protection creates more value than it gives away.

Trade secrets can be just as valuable

For many AI startups, trade secrets are often the most practical form of protection. Unlike patents, trade secrets do not require public disclosure. Instead, protection is maintained through confidentiality and controlled access. This approach can be especially valuable for:

• Training data
• Proprietary datasets
• Model architectures
• Validation methods
• Internal workflows
• Business intelligence

The speakers highlighted that if a competitor cannot easily reverse engineer a solution, keeping it secret may provide stronger long-term protection than filing a patent. This is where confidentiality agreements, NDAs, internal security measures, and carefully drafted contracts become essential.

Copyright protects code, but only to a point

Software source code is automatically protected by copyright. However, copyright only protects the specific code that has been written. It does not protect the underlying functionality or idea behind the software. As AI-assisted coding tools become increasingly sophisticated, competitors can often recreate similar functionality without copying the original code. For that reason, startup founders should not rely solely on copyright protection. Combining copyright with contractual safeguards and technical protections is often a more effective strategy.

The myth of data ownership

One of the most interesting discussions focused on data. Many AI companies view their datasets as their most valuable asset and assume they “own” the data. Legally, however, data ownership does not exist in the same way as ownership of physical property. Instead, companies protect data through:

• Licensing agreements
• Usage restrictions
• API access controls
• Technical safeguards
• Contractual obligations

This distinction becomes particularly important for startups building products around large datasets or providing data-driven services.

Protecting your brand matters too

While technical protection often receives the most attention, the speakers strongly encouraged founders to think about their market-facing assets as well. Trademark protection helps startups defend their company name, product names, and market reputation. Registering a trademark can provide significant advantages, including:

• Preventing copycats from using similar names
• Simplifying domain name disputes
• Accelerating takedowns of fake social media accounts
• Protecting long-term brand value

The recommendation was straightforward: every startup should consider registering its core brand name as a trademark early in its journey.

An overlooked tool: Design protection

Design rights are often underutilised, particularly by software companies. Unlike patents, design rights protect appearance rather than functionality. For digital products, this can include:

• Graphical user interfaces (GUIs)
• App layouts
• Visual product elements
• User-facing design features

For startups competing heavily on user experience and product design, design protection can provide an additional layer of defensibility at relatively low cost.

Final thoughts

The webinar concluded with an important reminder: intellectual property rights are tools, not goals. Successful startups do not necessarily have the largest patent portfolio. They understand where their value lies and select the protection mechanisms that best support their business objectives. For AI and data-driven startups, that may mean patents. It may mean trade secrets. It may mean trademarks, contracts, or a combination of all of them.

The key is to identify what truly creates value and build a protection strategy around it. As the speakers emphasised throughout the session, the question is not whether you should use intellectual property. The question is how intellectual property can best support your startup’s growth.

Access the full webinar replay in the Swiss Startup Association Education Library, free for members. Not a member yet? Join the community and get access to practical sessions that help you protect your business before something goes wrong. 

Don’t miss out on the latest news and events. Subscribe to our newsletter and stay up to date.