Chile Case study

The National Institute of Industrial Property (INAPI) are committed to delivering the highest quality service to their public. With the help of TrademarkVision, examiners have fast, accurate results to facilitate the decision-making process


Chile’s National Institute of Industrial Property

Chile joined the TrademarkVision revolution and became the first South American country to integrate image recognition and artificial intelligence in their examination process. We’re delighted to be working with another government IP office dedicated to embracing innovative solutions for industry change.

Explain the problem the INAPI was facing with image searches before TrademarkVision?

One of the most time-consuming steps is finding relevant trademarks consisting of figurative signs. Vienne codes are not enough and the offices must allocate enormous resources to classify figurative elements and conduct searches that regularly find thousands results, with no prioritization criteria among them. 

Our examiners then spend a lot of their scarce time reviewing those results, slowing the process and allowing a high chance of human error. This naturally might affect the quality of our work.


Why is image search important for your public?

About a 50% of trademark applications contain figurative marks. Image search is important as the quality of substantial examination is improved, so will the predictability of INAPI’s decisions. Thus, giving users a clearer idea of the chances of refusal for a trademark application and facilitating a better decision making process. 

It will also give users a comprehensive understanding of the reasons why an application has been denied and perhaps point them in the right direction regarding changes they need to make to their branding, in order to obtain protection via trademark registration.

How does TrademarkVision help your process?

TrademarkVision allows us to conduct image searches during the examination process.

The results are available in seconds and the most relevant ones are placed at the beginning of the list. This has an enormous impact on the quality of trademark examination, as we are no longer partially blind when it comes to figurative signs. We can now base a refusal decision on a solid reason, not only due to the same Vienne codes but the comparison of the images themselves.

It is quite intuitive and our examiners were ready to use the tool after one single training session.


What other areas do you think will be able to leverage AI in the IP space?

It would be interesting to see an AI based system that could automatically perform an examination of goods and services according to the Nice Classification and other international lists of goods and services, such as TM5’s IDList. 

As of today, a large part of the examination in this matter is performed manually by classification experts within our office, which undoubtedly takes a lot of time and occasionally generates mistakes or inconsistencies.