Updates on IPrate coverage, data, and methodology.
This release covers 31 European countries: 28,219 ranked firm profiles (+293 vs previous release) and 12,064 ranked attorney profiles (+66 vs previous release), computed from 18,617,957 official filing records at EUIPO, EPO and national IP offices. Ratings are recomputed in full with every release and are never sold or sponsored.
This release covers 31 European countries: 27,926 ranked firm profiles and 11,998 ranked attorney profiles, computed from 18,465,526 official filing records at EUIPO, EPO and national IP offices. Ratings are recomputed in full with every release and are never sold or sponsored.
At first glance, the EUIPO's updated guidance on the responsible use of generative AI seems like a welcome signal for AI governance. Transparency, provenance metadata, data classification and human oversight are foundational principles. They are also cited by almost everyone who wants to appear in control of technology that is moving faster than legal rules. Yet that is precisely the problem.
In trademark examination, AI is more than a productivity tool. It can influence how conflicts are identified, how similarity is framed, how evidence is summarised, and how risk is communicated to applicants. Even when AI does not decide, it can still shape the decision environment. A weak signal, badly presented, can become a strong assumption in the mind of an examiner, lawyer or founder. A human reviewer who is rushed, undertrained or over-reliant on machine output may simply become the final rubber stamp for AI. Human oversight is not a magic shield.
The emphasis on provenance metadata raises other hard questions. What exactly will be recorded? The source documents? The model version? The prompts? The retrieval logic? The confidence thresholds? The human edits? Provenance should not be a vague label; it should be a meaningful audit trail. Users need to know what was generated, what was retrieved, what was inferred, and what was legally verified.
This is especially important in legal and intellectual property workflows, where faster decisions are only useful if they are reliable, explainable and predictable. AI systems should expose uncertainty, preserve source links, separate prediction from legal conclusion, and make it easy for users to challenge the output. Data quality is not a technical detail. It is the foundation of trust.
That is also the standard IPrate is built around. IPrate ratings are data-based and rely on deterministic criteria rather than subjective assessments or unsupported inferences. The goal is not to replace expert judgment, but to provide structured, repeatable signals that can support better decision-making.
The EUIPO guidance is a step forward. But the real test is whether AI automation in trademark practice will be balanced with accuracy, accountability, reliability and repeatability.
The European Patent Office (EPO) has published the 2026 Patent Filing Guidelines, which entered into force on 1 April. These rules bring together guidance that was previously spread across different parts of the European patent system. The EPO is trying to make its procedures clearer and more consistent for applicants who file and manage patent applications in Europe. Entrepreneurs, researchers and businesses dealing with innovation should understand that the Guidelines shape what patents actually protect in Europe and which innovations qualify for patent protection.
One practically important change is that European patent applications can now be filed in DOCX format. DOCX files are easier to edit than PDFs for non-technical users. This may make it easier for inventors and rightsholders (especially startups and small and medium-sized enterprises) to file patent applications themselves without involving professional patent firms. It may also lower barriers to entry for new patent firms, which no longer need sophisticated electronic filing systems. For established players, however, it may increase complexity, as changes will need to be integrated into existing electronic filing systems.
Due to the complexity of international patenting, anyone thinking about filing for patent protection with the EPO is strongly advised to consult a qualified patent professional.
IPrate is now available in public beta, offering the first fully data-driven ratings of intellectual property firms across the European Union and European Economic Area. The ratings cover trademarks, designs, and patents, and are intended to fill a long-standing gap: the absence of objective, transparent, and independent performance information in the European IP services market. Until now, the available ratings have relied on self-reported surveys, peer nominations, and paid editorial placements — none of which reflect actual filing track records.
IPrate firm ratings are built entirely from official filing and outcome data published by the European Union Intellectual Property Office (EUIPO) and the European Patent Office (EPO). The methodology evaluates firms across six measurable dimensions — volume, success rate, procedural efficiency, opposition survival, renewal commitment, and domestic client trust — and applies Bayesian credibility adjustment to ensure that scores reflect genuine evidence rather than statistical noise. During this public beta, all ratings are computed at the European (Euro) tier, comparing firms based on their activity before these two pan-European offices.
National-jurisdiction data — covering filings at individual country patent and trademark offices — will be integrated in the near future, adding a domestic tier alongside the existing European view. This will allow users to see how firms perform not only at the continental level but also within their home markets. Country-by-country rollout is already underway, and updates will be announced on this page as new jurisdictions come online.
