AI is reshaping the e-commerce landscape – improving recommendation engines, streamlining logistics, personalising user experiences, and automating everything from pricing to customer support. With this rapid growth comes a wave of innovation and protecting that innovation is key to staying competitive. Whether you’re a startup deploying cutting-edge AI tools or an established retailer enhancing your online platform, it’s essential to navigate the IP landscape strategically.
Here are 5 top tips to consider when pursing protection for AI inventions in e-commerce:
-
Technical improvement over existing technology
Patent law in some jurisdictions excludes “methods of doing business” from protection unless they demonstrate a technical effect beyond improving the operational efficiency or profitability of the business. E-commerce solutions – like AI-based recommendation systems or dynamic pricing methods – can fall into this grey area.
To strengthen your chances of obtaining a patent, focus on what technical contribution your invention makes over existing systems. Is it improving computational efficiency? Does it enhance data security or enable faster processing at scale? The more clearly you can articulate these benefits in technical terms, the better positioned you’ll be to overcome business method exclusions.
- IP options: patent or trade secret?
Not every innovation is best protected with a patent. While patents offer strong protection and potential commercial leverage, they do require public disclosure – and that’s not always ideal when it comes to AI algorithms. For AI inventions where the value lies in a proprietary training dataset, a unique training process or model weights that are hard to reverse-engineer, trade secrets may be a better form of protection.
However, it’s important to be aware that trade secrets offer no protection if someone independently develops the same invention. Therefore, consider your business model, your competition, and how easily your solution could be reverse-engineered when deciding between patent and trade secret protection.
- Acceleration options
AI evolves quickly, and so do your competitors. Traditional patent timelines may seem comparatively slow and struggle to keep pace with the market. Fortunately, many jurisdictions offer accelerated prosecution routes.
The EPO’s PACE program can be used to accelerate the search and examinations stages of prosecution. Beneficially, the PACE program has no associated official fee, however it is important to respond to all communications in a timely manner as any requests for deadline extensions will lead to removal from the PACE program. The UKIPO’s Green channel also offers accelerated prosecution for inventions with a clear environmental benefit. The request can be made at any time of prosecution with a statement on the environmental benefits of the invention e.g. operating more efficiently and requiring less energy. This method also beneficially has no associated official fees.
These programs can help get your patent granted faster – potentially giving you enforceable rights when they are most needed. You can ask your attorney to discuss the acceleration options in the jurisdictions you are interested in and whether your application qualifies.
- Sufficiency of your application
Patent applications must disclose inventions in sufficient detail to allow a skilled person to reproduce them. With AI inventions, this can be tricky – especially when the innovation lies in a complex, trained model.
If your patent application glosses over how the model was trained or what data it used, it may fail the sufficiency requirement. Be prepared to include enough detail to explain how your AI system works, how it is trained, and what technical advantages it provides. Keep in mind, this may affect how much of your innovation you are comfortable disclosing which reinforces the trade-off between patent and trade secret protection.
- First filing options
When it comes to AI-based inventions, particularly in e-commerce, the likelihood of securing grant can differ between the UKIPO and the EPO. This divergence stems from the differences in how each office assesses inventive step, technical effect and exclusions relating to computer programs ‘as such’.
The EPO uses a two-hurdle approach, focusing first on whether the invention solves a technical problem with technical means i.e. any hardware – this step is normally straightforward to pass. It can then proceed to assessment of inventive step, where non-technical features may be considered in combination with technical ones if they contribute to the solving of a technical problem. This allows AI inventions that improve system operation, processing efficiency, or data handling to stand a better chance.
By contrast the UKIPO applies a different test derived from the Aerotel/Macrossan judgement to assess excluded subject matter before any consideration of novelty or inventive step, which can lead to broader exclusions for business-related or computer-implemented inventions. Under this approach, many AI-based innovations – especially those centred on business logic or user interaction – may be refused on the basis of being excluded subject matter.
For this reason, filing first at the EPO may offer a stronger path to grant, particularly for AI software innovations. Beneficially, a patent granted by the EPO can still offer protection in the UK.
(Note that the resulting differences in patentability using these two approaches is currently being considered in the Emotional Perception UK Supreme Court case we discussed here.)
AI inventions in e-commerce sit at the cutting edge of innovation – but also often at the limits of patentability. By carefully considering the above points, you can maximise the value and protection of your innovation. Get in touch with one of our expert attorneys in this field to ensure your IP strategy aligns with your invention and your commercial goals. In the fast-moving world of AI and e-commerce, IP can be one of your most valuable assets.
.png)
.png)