20+ years of historical data
Training set built from over two decades of INPI decisions — approvals, rejections, oppositions, and appeals. Each dispatch published in the RPI feeds the model.
Proprietary Technology
TrademarkIQ Predict®
Anticipate the risk of rejection before filing.
Robust statistical process, trained by machine learning and artificial intelligence technology, with over 20 years of historical INPI data and 100 million data points analyzed. An unparalleled technology in the world for intellectual property market intelligence.
Predict®
proprietary score
Calibrated on real INPI decisions, validated through back-testing
20+
years of historical data
Since the first digital records of INPI
100M+
data points analyzed
Bibliographic, procedural, and classification data
Weekly
update cycle
Each RPI published feeds the model
How it works
A model built on real data, not assumptions.
Predict® is not a static checklist. It is a living system, fueled by every INPI decision since the first digital records — and it evolves every week.
100M+ data points analyzed
We cross-reference bibliographic, classification, procedural, and temporal data to build a multidimensional profile of each trademark filed in Brazil.
Robust statistical process
The Predict® score is the result of a rigorous statistical process, trained by machine learning and artificial intelligence technology. It is not a simple heuristic — it is a model calibrated on millions of real observations.
Continuous evolution
Each new Industrial Property Review published by INPI feeds back into the model. Predict® continuously improves — the more data, the greater the accuracy and granularity of the predictions.
Model robustness
Accuracy that stands up to scrutiny.
Predict® does not rely on fixed rules or keywords. The model learns complex patterns from millions of real decisions — and is continuously recalibrated as new RPIs are published.
This means that the score reflects the current reality of INPI, not a static snapshot of the past. Every week, the model incorporates new decisions, new examination patterns, and new approval trends.
Continuous calibration
Each RPI published by INPI feeds back into the model. The training set grows weekly with new dispatches, approvals, and rejections.
Multidimensional
The score is not a single metric. It crosses phonetic, semantic, classification, temporal, and procedural dimensions for each prediction.
Validated against history
The model is rigorously back-tested against decisions already issued by INPI — performance measured in real data, not estimates.
Applications
From individual assessment to portfolio strategy.
Before filing
Know the probability of approval before investing time and money in the filing. The Predict® score anticipates obstacles that a simple search does not reveal.
Portfolio management
Evaluate the aggregated risk of each client's trademark portfolio. Identify vulnerable marks and prioritize actions based on data — not intuition.
Competitive Intelligence
Analyze the probability of competitors' trademark applications being granted. Anticipate moves and make strategic decisions with a unique perspective that no one else offers.
Ready to decide with data?