The difference is huge

Find Landscape Insights a human eye cannot see.

A Quantum Leap

Find Hidden Technologies to Invalidate

A Paradigm Shift

Fixing the broken system of Patent Search

A Cut above Rest

Specialized Technology Standard Mapping System

Difference between ordinary
and extraordinary results

Welcome NEO, a deep learning system developed by GreyB.

Simply put, imagine the role of M. Jordan’s coach in his exceptional performance. That is exactly what NEO does.

It helps achieving peak performance in searching patents.

Imagine you found best prior art used key word “Control System” instead of “Transistor”. Seemingly unrelated words, but describing the same technology.

Remember, anyone can find related words – it is easier. NEO connects unrelated words describing the same technology.

Who do you bet on?

Jordan minus his coach or Jordan plus NEO.

History of Neo

Invalidity searches are laborious and frustrating as these require one to review 4000-5000 references manually to find ONE relevant reference that had powers to invalidate a patent; however, at the same time, prone to be skipped by human eye as it describes the concept in the most complex way possible and in a completely different manner.

And after running thousand of such validity studies, we decided to build an expeditious system. Our strong background in machine learning helped us conclude that computers, using the strong algorithms, could learn a thing or two from the brain and understand at a basic level how it manages information and process it, while analyzing patents.

The core approach to any deep learning problem is to represent the structure and constraints of the problem in a way that is computable & this is generally harder for some problems than others, and as we already see that there are certain areas where deep learning is already superior to a human. A system that can work with humans to deliver better performance than that of an individual. Our system analyzes the patent data-set to learn and discover relationships among entities in a similar way as humans try to do.

42 Million+ Megabytes of data

Neo covers more than 42 Million Megabytes of data while performing any analysis.

12000+ hours of development

Over 12000 hours of research and development time has been spent in building Neo to where it is now.

Protected by more than 10+ of patents

More than 10 patents have been filed, in United States & Singapore, to protect the technology behind Neo.

~5.6 Million of records ingested

About 5.6 Million of records already indexed and counting


It’s a methodology for learning high-level concepts about data, frequently through models that have multiple layers of non-linear transformations. It is about learning from what is already lying there. It is about learning about how one technology can be described in various ways by thousands of attorneys while filing patents.

Neo is a system that can predict the way in which a hidden prior art might have been described. If we talk about portfolio analysis, it can identify the hidden insight within the portfolio of a company’s patents. It can tell you where the technology is heading in the future. While that’s far off from the kind of AI found in science-fiction movies, but its close to a place where it is actually useful.

What can Neo do that others can’t?

There are various tools available in the market, that try to give you the results automatically, but the industry has rejected most of them as they were not accurate. A lot of them tried to generate automatic landscape reports, but failed miserably. Therefore, we put our heads to solve this problem and used our experience of working on thousands of invalidation searches to train a system that can think faster than a human. A system that can correlate 50 charts in 2 seconds while a human can only relate a maximum of 5 graphs at one go.

Discover what Matters

Neo can help find the hidden concepts that are described in a completely different manner and are difficult to be captured by any existing systems. Neo leverages the power of deep learning and contextual pattern matching. It is the first intelligent tool that can read patents as human.

How it works?

Till now patent tools were more artificial than intelligent. We ingested human brain power in the system to make it work intelligently with patents.

For example, through permutation and combination, Neo has the capability and knowledge to apply 1.35 million logics on the patent data to derive insights which may be difficult to formulate by a human researcher. Landscape projects can now be executed in a completely different way.

Churning more than a human brain?

Neo has gone through millions of patents and built its understanding by studying the conceptual relationships on how the technologies have been described. So Neo can build a timeline describing the different ways how a particular technology has been described in last 50 years. It has built the capability to link expected and unexpected connections between various technology descriptions.

How does it understand patents?

Neo is the first and the only Neuro-Linguistic platform that uses the true power of the human brain for the analysis of patents. We can separate the information from the intelligence. After executing thousands of searches, we could see a pattern develop whereby a human is able to spot a difference, but a tool misses. With our strong interest in Artificial Intelligence, we used this human capability and rooted it into an engine to build a deep learning capability.

Got queries. Fill in your contact details and we'll get back to you ASAP.

Become a part of GreyB’s insider list

Get our distilled learning delivered to you.

Get the Sample Report

Fill out the form and get the report.