Through my career as an IP consultant, advising companies on how best to develop an IP strategy to create and grow value through IP assets, I have often seen software businesses relying on copyright for IP protection. Patents were often seen as too expensive and for hardware type business and not relevant to business innovating faster that than the time it takes to get a patent. Trade secret protection was usually seen as too bureaucratic and against the internal sharing culture encouraged within software companies. The tide may be turning.

AI coding agents are redefining software engineering. They have the ability to generate software code autonomously by “translating” user intent into code. This has led recent releases of the latest AI models to impact the share price of multiple SaaS businesses.

How can software SaaS companies protect themselves from being replicated through AI generated code with a strong IP strategy?

This is a new era for both software engineers and IP strategists where new thinking is required. Below are some potential components of a more holistic IP protection strategy in the age of AI coding agents. These will need to be reviewed constantly given the pace of development of AI models and potential regulatory changes.

1. File patents: The rules vary from country to country, but software-related inventions are generally patentable if they have “technical character” (e.g. offering more than administrative or aesthetic benefits). An experienced patent attorney can help to assess the likelihood of obtaining patent protection for given software products. Even in cases where the prospect of obtaining broad patent protection is lower, having a pending patent application in place can still provide a useful deterrent effect.
2. Implement a robust Trade Secret strategy: a multi-angles strategy would be required to shield the codebase from the end-user, hence also from the AI coding agents. It would involve internal security, AI usage and HR policies to avoid leak of the code. It would certainly require to build a contractual protection layer through End-User License Agreement (EULA) restricting reverse engineering and scraping and unauthorised data extraction. An additional measure would be to hide core logic through running the main logic on private servers while the user’s machine operates the interface and display the results. Other technical measures such as code obfuscation or digital watermarking would increase the protection of the trade secrets.
3. Build a moat and leverage proprietary data sets: data accumulated through years of production, customer dealings, R&D, and other business operations form valuable data sets. Producing insights from tapping into your own private data act as a differentiator that may not be easily replicated through AI coding agent.

As AI is dramatically impacting SaaS businesses, developing a robust IP protection strategy will require not only to use new measures but also to connect multiple functions and departments within the business such as legal, HR and engineering who are not always accustomed to work together. IP advisory firms can act as an interface.