This post is also available in: heעברית (Hebrew)

As opposed to data, knowledge has no shape. It is chaotic, constantly changing, and unstructured in nature. Correspondingly, while there are available technologies that enable the efficient analysis of structured data, there is still a huge technological gap in the extraction of value hidden in raw data and all aspects of soft, unstructured data (such as free text).

In an era of information explosion, vast amounts of soft textual data are produced ubiquitously and make insight extraction and knowledge refinement near impossible. As a result, the knowledge gap and loss of potential value are ever growing. 

In order to try and meet current knowledge gaps and the loss of value, organizations today are required to invest vast amounts of resources in the process of fusion and analysis of textual big-data.

The value extraction process can be compared to oil searching or gold mining:  frequently, the process is not economical due to the large budgets required, while the value hidden in raw data can only be unveiled in retrospect.

This phenomenon is especially evident in complex languages such as the Semitic and Asian languages, where current natural language processing (NLP) technologies struggle to cope.

In order to solve this problem, Israeli startup KEOTIC has developed a ground-breaking artificial intelligence (AI) technology that yields structure into raw unstructured data by connecting untrivial dots across data-silos and understanding the “bigger picture”, making processes that couldn’t have been performed to date – a reality.

KEOTIC has adopted a revolutionary approach for information fusion and extraction from large and complex textual databases.

The company has developed CognitiveCore AI Engine – the first AI engine for knowledge extraction imitating the human memorizing and learning process.

The AI reads raw unstructured data and processes it, just as the human brain does, by connecting themes and learning the ‘rules of the world’. This enables capabilities such as semantic search, prediction, advanced analytics, and the uncovering of unknown-unknowns, all performed in a fully automated manner, in every language and any given content domain.

Government and law enforcement agencies have been the first to acknowledge and examine the unique value offered by KEOTIC. 

In a POC conducted with one of Israel’s security arms, Keotic’s AI Engine’s capabilities were proven just hours after the system was fed with hundreds of thousands of documents. CognitiveCore’s capabilities were proven useful and provided tremendous value not only in flaws prevention and prediction but also in safeguarding human lives. 

KEOTIC’s AI managed to locate and connect non-trivial information dots, revealed unknown-unknowns, found “needles in a haystack” and enabled additional capabilities including semantic search and predictions. 

How does the technology work?  The AI understands the significance and context of the given information pieces without focusing on the individual word level or even the textual representation itself. instead, the AI-based system operates parallel to human properties, and according to the vastest and deepest taxonomy that it can find in the given data. 

The system’s agile installation on-premise or on the cloud constitutes a considerable advantage for organizations with classified data as KEOTIC does not need to see any data. Following a simple installation and a short orientation phase, the client is able to operate the system independently. This obviously constitutes a huge advantage for the HLS realm – police, intelligence, and other agencies.

KEOTIC’s solution provides apparent business advantages:

  • Redundancy of the prolonged work of scarce, and expensive, Machine Learning and NLP experts.
  • The system is fully automated and is agnostic to verticals (content subject) and languages, including the most morphologically-rich languages such as Hebrew, Arabic, Chinese, etc.
  • Huge savings in resources as the system provides value right on the day the text database is accepted. The company has defined unique service level factors, eg 1D data2value SLA. 
  • The ability to locate and extract unknown-unknowns – clients don’t have to know in advance what they’re looking for. 
  • Agile installation.

In 2020, KEOTIC operated within the INNOFENSE Innovation Center – a unique program for startups in security fields, led by iHLS in collaboration with Israel DDR&D (MAF’AT), and Israel’s Ministry of Defense.  

Roni Wiener, KEOTIC’s CEO and founder: “Participation in the program has enabled us to get acquainted with, and penetrate, the HLS market, where we witnessed an enormous need for our technology. Although our efforts were initially directed towards the civilian market, INNOFENSE has efficiently opened doors for us to realms with very high potential”.

What are the company’s plans for the longer term?

“We believe that in the near future AI systems for information management and extraction will become common in organizations the same way emails and coffee machines are. We focus the development of our technology and products in this direction and concentrate on producing that “organizational-brain” that reads, memorizes, and unifies all current and historical information in and outside the organization”, Wiener reveals.

“That same “organizational-brain” will proactively assist in decision-making processes, as well as in problems and opportunities exposure. It operates in an independent, proactive manner without the need for explicit requests from users. Organizations and companies that will be late to adopt such technologies will be left behind.”

In a world full of data and information, KEOTIC undoubtedly makes sense in chaos, providing organizations with immediate and unprecedented value through its proprietary, fully automatic solution.