Infringement Detection

Infringement detection answers one question: which products on the market practice the claims of a patent you hold? You start from your own patent and look outward, scanning competitors for products that read on its claims. This is the offensive direction. Your patents sit on one side, the suspect products on the other, and a mapping between each pair scores the overlap.

The work is to turn a crowded market into a short list: the few products whose features actually practice your claims, ranked by the evidence behind each one. Those are the cases worth investigating, licensing, or asserting.

You run it from the Offensive / Infringement Detection workflow, which spins up a new research database built around your assertion. It asks which patents you are asserting, seeds a search for products that might infringe them, and lays out the pages you will read the results on. Keep one patent family per research so each verdict stays clean. The three steps below take it from your patents to a ranked shortlist of suspects.

1. Choose patents#

Bring in the patents at the center of the case. The claims are what the rest of the analysis works against, so this is the side you want exactly right.

There are two ways to add patents:

• Enter the publication numbers directly when you already know the family you are asserting.
• Build a query against the global corpus by keyword, semantic meaning, classification, applicant, or date to round out the portfolio.

The numbers appear at once, and the full records (claims, families, classifications) fill in as the corpus is read.

2. Search for products#

Now define the other side: the market products to screen against your claims. These are other companies' offerings, profiled from the outside, so the routes differ from freedom-to-operate, where you upload your own datasheet. Build the set with a products filter and populate it two ways:

• Search the product corpus. It finds products by their features and category. Point it at a competitor or a product class and Patalyze surfaces matching items from the wider catalog, so you can pull in everyone who sits in the same space as your invention.
• Crawl from a website. It fetches product information straight from one or more URLs. A competitor's product page, spec sheet, or datasheet on the open web becomes a product in your database, with its features read out automatically.

From each product's description, files, and images, Patalyze learns its features in the background. Those features are exactly what each claim element gets compared against, so the more of a competitor's source material you gather, the sharper the read.

3. Rank the results#

With your patents on one side and suspect products on the other, Patalyze creates the mappings automatically. For each patent-and-product pair it breaks every independent claim into its elements, judges whether the product practices each one, and rolls the verdicts into a per-claim score and a risk band.

There are thousands of them, and you will not read them all. Three surfaces turn them into something you can act on, from the headline down to the evidence:

• Read the executive summary. Patalyze fronts the research with a generated note that states the verdict in plain language, with live links back to the patents and the products that read on them. Start here for the short version before you open the rows.
• Open the products table. Each row in the table is a suspect product, and every patent you are asserting adds a column of claim scores. Filter those columns to the critical and high bands and the market collapses to the few products whose features actually practice your claims.
• Inspect the mappings. Every score in that table is a mapping. Open a cell to see each claim element, the verdict and evidence behind it, and the risk band they roll up to. A Critical or High band on a patent you hold and have in force is an infringement candidate to investigate or assert. You can also map by hand and override a verdict you have ruled on yourself.

To see the shape of the field instead of the rows, add a dashboard. It breaks the matches down by manufacturer, category, or territory, so you can tell which competitors and markets cluster the risk.

New here? The quickstart walks the same path from the first click.