Elemendar was founded in 2017 by Giorgos Giorgopoulos and Syra Marshall at the UK’s first GCHQ / NCSC Cyber Accelerator, powered by Wayra UK, to develop Cyber Threat Intelligence (CTI) enrichment capabilities.
Elemendar is the leader in developing AI to translate CTI into machine readable and instantly actionable data. This can be fed into systems such as SIEMs and TIPs directly to reduce risk and return valuable time to analysts at the forefront of organisations’ defence across enterprise, government and law enforcement. Elemendar’s technology is used by both government and private customers.
Our AI automates the reading and translation of CTI from human authored unstructured text and documents into machine readable and actionable data output as STIX 2.0 and incorporating MITRE ATT&CK™. Our technology utilises leading-edge Machine Learning (ML) and Natural Language Processing (NLP) technologies.
We have created Elemendar as we are passionate about security. Organisations and cyber analysts are drowning in critical CTI and we want to make it usable for all, not only the 1% who can afford analyst teams.
Dr Kimmo Soramäki
What We Do
Elemendar translates unstructured reports into STIX outputs for Cyber Analysts.
Our AI makes your analysts more efficient, instantly flagging threats.
We read reports in seconds, not hours.
Save your analysts’ time.
Empower them to focus on what’s most important.
OUR CURRENT TECHNOLOGY
In one hour of an analyst’s job, 50 minutes is reading, 10 minutes is valuable analysis. Elemendar shortens the reading to 50 seconds, increasing efficiency up to 5x. Our customers using the A.I. say it can perform as well as a junior analyst.
Incoming C.T.I. is automatically converted into machine readable, actionable data. Using STIX2 (and imminently, MITRE ATT&CK™), this can be exported / imported directly into a SIEM or TIP.
Elemendar uses long short-term memory (LSTM) recurrent neural networks (RNNs), building on NeuroNER, Tensorflow and spaCy. We currently achieve precision >98% and recall >91% for NER. As our A.I. learns more, this will go even higher.
Where We’re Going
C.T.I. is read and translated into STIX2.
Networks are built by correlating across source C.T.I. docs.
Relationships are understood between multiple data points. Network analysis to find patterns
Patterns are used to share enriched context with other tools going both ways. Iterative machine-led action on C.T.I.