The patent system has seen a large number of software patents invalidated over almost a decade since the Abstract Idea 101 analysis was created. While the pendulum has swung back in favor of patent owners in recent years, there are important steps that companies can take to drastically increase the validity of their software patents.
In 2014, the Supreme Court of the United States (SCOTUS) decided Alice Corp v. CLS Bank Intl. creating a new doctrine for invalidating patents as directed to an abstract idea under 35 U.S.C. § 101. In its essence, the new doctrine prevents existing processes and methods from simply being performed by a generic computing device (at least in the software technical field). As an example, creating a video game to play the game of chess was found unpatentable because the game merely took an existing board game and used a generic computing device to present and manage the game of chess. However, in my experience, invalidated patents tend to share one thing in common - they lack technical details, usually because the patents were prepared with very low budgets, by practitioners that failed to understand the technology, or by practitioners that were not provided with enough technical details when preparing the patent application. Once the patent application is filed, it is too late to add any additional technical details.
To prepare a "valid" software patent, the most important steps are to 1) frame the invention from a technical perspective and 2) spend the time to include the technical details. For example, rather than the patent application focusing on a video game that implements the existing game of chess, the patent application should be framed to solve the technical problems of implementing such a game, including details such as how to build and train the artificial intelligence model that scales to fit a player's skill level. Further, it isn't enough to include a single sentence in the specification stating to train the artificial intelligence model. Instead, the patent application should include technical details on the steps to train the model including steps to identify and correct errors, steps for retraining on an ongoing basis, details on input variables used in the training process, and details on how the training data is captured, processed, and stored.
As a takeaway, it is important to realize that all patents are not created equally. To get the most value out of your patents, find a patent attorney that clearly understands your technology. Also, if you need to cut costs, do it somewhere other than in the drafting of the patent application. There are numerous ways to reduce or delay costs during the patent process (see, here for example). The drafting of the patent application is the wrong place to save money if you want an enforceable software patent.
