Artificial intelligence based system and method for intelligent ranking of it assets and predicting enhanced cyber-attack surface for an enterprise
An AI-based system utilizing multiple machine learning algorithms ranks IT assets to provide comprehensive visibility into the attack surface, addressing the complexity of IT environments and enhancing cybersecurity decision-making and threat management in enterprises.
Patent Information
- Application Number
- PCT/IN2023/050830
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- Priority Date
- 2023-05-06
- Filing Date
- 2023-09-02
- Publication Date
- 2024-11-14
AI Technical Summary
Complex IT environments in enterprises lack end-to-end visibility, making it difficult for IT professionals to take timely and accurate decisions, and traditional cybersecurity tools struggle to handle the volume and speed of data required for effective threat detection and response, especially given the limitations of single ML algorithms and incomplete AI training data.
An AI-based system using a combination of supervised and unsupervised machine learning algorithms to rank IT assets based on their criticality, interdependencies, and usage, providing a comprehensive and dynamic view of the attack surface, enabling proactive, detective, and remedial measures.
This solution enhances the ability of IT professionals to prioritize vulnerabilities, direct resources effectively, and respond to cyber threats by providing accurate and timely insights into the attack surface and security posture of enterprises, improving decision-making and threat management.
Smart Images

Figure IN2023050830_14112024_PF_FP_ABST
Abstract
Description
BACKGROUND
[0001] Today’s businesses are highly dependent on Information Technology (IT) infrastructure. In big enterprises there are many variables that makes IT environment very complex. Some of the variables are hard to identify and measure and are connected in a way that are complex to understand. Also lack of end-to-end visibility shadows IT professionals’ ability to take timely and accurate decisions.
[0002] Threat detection and response is one of the most important aspects of cybersecurity but due to the enormous volume of data generated and lacking in speed of execution, traditional cyber security tools are facing too many obstacles to apply this approach.
[0003] There are innovative work and systems developed using Artificial Intelligence (Al) models to handle large volume of data and increasing speed of execution, however, those primarily concentrate on sub-part of IT environments (e.g., IT network, End users, Databases etc.) or that try to detect anomalies in the IT operations to predict a cyber-attack or a security incident. Most companies today have extremely complex infrastructure that are made up of applications, systems, and networks in environments consisting of on-premises data centres, public and private clouds and these inventions are unable to protect the entire threat surface.Most of such models lacks the Al training data for these models to make it more effective. Also, use of one type of Machine Learning (ML) algorithm at a time has limitations to handle rapidly evolving cyber-attack landscape.Therefore, these approaches have not been very successful yet.
[0004] To remediate cyber security threats and vulnerabilities, organizations need clear visibility into their attack surface, the critical assets within it, and the threats and risks to those assets. With this visibility, security teams can perform a cyber security vulnerability assessment, prioritize the most critical vulnerabilities, and direct resources to remediate the greatest risks.SUMMARY OF INVENTION
[0005] Managing the risk posed by ineffective cybersecurity protections is an ongoing part of all business operations. The threat landscape is always changing, new exploits are discovered and new potentially vulnerable devices that increase the attack surface are frequently added to the network. Therefore, ensuring that cybersecurity protections are maintained over time is essential. Discovering and mapping all digital assets to quantify the attack surface is vital for organizations to prioritize possible threats and minimizing their attack surface.
[0006] Our invention primarily focuses on Ranking of IT assets for an enterprise, interoperability and inter-dependencies of IT assets so that criticality of each IT assets assigned as per weightage. This ranking can then further used for proactive, detective and remedial measures such as:Determining the cyber-attack surface in an enterpriseDetermine security posture of different Business applicationsAssist IT security professionals to prioritize proactive actionsIn case of abnormal behaviour or incidents provide visibility on other IT assets to investigate forThreat modelling
[0007] Our invention is primarily based on the data available in the enterprise where it will be used. Make use of multiple ML algorithms and combination of Al models (Supervised / unsupervised) to produce the results more accurate and consistent.
[0008] Though the current focus is to use this system and method in Cyber Security field, this can be used in any field by using similar hypothesis as applicable.The Hypothesis and Features of the design provided herein can be better understood with reference to Drawings, Descriptions and claims, all of which form the disclosure of this patent application.BRIEF DESCRIPTION OF DRAWINGSThe Drawings refer to some embodiment of the design provided herein in which:
[0009] Drawing 1 depicts a tiny representation of complexity of IT Infrastructure deployed in an enterprise creating a wide attack surface and making IT professionals’ job very difficult to take timely and informed decisions.
[0010] Drawing 2 illustrates a block schematic of the hypothesis and system where Al model takes inputs from the different sources including human analysis, processes data on trained supervised and unsupervised Machine Learning Algorithms to produce intelligent assets ranking and enhanced prediction of attack surface with rational so that IT professionals can take informed decision and actions.
[0011] Drawing 3 describes the steps in formation of the hypothesis and system of the design.
[0012] Drawing 4 illustrates the block diagram of an embodiment of the Al system that predicts enhanced predication of the cyber-attack surface that can help for prioritizing and taking informed “protection measures”.
[0013] Drawing 5 illustrates the block diagram of an embodiment of the Al system that support or refute abnormal patterns observed in the IT assets for enhanced “detective measures”.
[0014] Drawing 6 illustrates the block diagram of an embodiment of the Al system that can help for prioritizing and taking informed “remediation measures”.
[0015] Drawing 7 illustrates the block diagram of an embodiment of the Al system that can help IT professionals to derive “security posture of individual business applications” and “Threat Models”.DETAILED DESCRIPTION
[0016] In the following description, specific details are set forth and explained with respect to the drawings in such a way that an ordinary skilled persons can practice the design without these specific details.While the design is subject to various modifications, equivalents and alternative forms, specific embodiments are shown by the way of example in the drawings and will not be described in detail. It should be understood that the designs are not limited to the particular embodiments disclosed, but on the contrary, the intention is to cover all modifications, equivalents and alternative forms using the specific embodiments.
[0017] Today’s businesses are highly dependent on Information Technology (IT) infrastructure. In big enterprises there are many variables that makes IT environment very complex. Some of the variables are hard to identify and measure and are connected in a way that are hard to understand. Also lack of end-to-end visibility shadows IT professionals’ ability to take timely and accurate decisions.
[0018] As depicted in Drawing 1, IT Infrastructure environment in an enterprise comprises of different network segments, different operating environments (105, 107, 108, 109, 110), virtual infrastructures (111) that shares resources among different applications, cloud computing environments external to enterprise (102), access to employees, external users (101) and third-party vendors (101) etc.Below are some of the scenarios explained that this system and method is trying to address.1. Criticality of the IT assets to an Enterprise: Same type of IT assets are used for different purposes, different operating environments and different business applications. So same type of IT assets has different criticality.For e.g.• A firewall used to separate non-production & production networks within Data centre (106) and a firewall used to separate enterprise network and outside world (104) has different criticality ranking for the organization• A web server used internally for employees (107) and the one places in DMZ (De-Militarized Zone) (105) for outside world has different cyber security implications• A database (109) used for core business application is more critical than to that of used in other supporting applications or non-production environments (110)2. Sharing of IT resources: Due to increased use of Virtualization, IT assets (111) are shared among different business applications causing inter-application dependencies and threats that are difficult to analyse.For e.g.• There can be many virtual Windows servers used for different applications hosted on single Physical server• Multiple databases or layered applications configured on single server3. Lack of visibility of the usage: Due to increased use of cloud hosted environments (102), access to enterprise IT resources for 3rdparties (103) or integration with other supply chain partners, it is very difficult to understand if components of critically ranked assets are accessed by different entities or users and assess the vulnerability or threats from them.Therefore, criticality and security measures apply based on the placement or usage of the IT asset and a unified principle cannot be applied IT asset (even if it is same type of asset) in any enterprise. This requires determining the Attack Surface for an organization.
[0019] Traditional legacy tools lack the correlation intelligence and fail to handle overwhelming amount of data that needs to be processed. Therefore, modern threats bypass these tools on daily basis and for humans it is virtually impossible to scrutinise this amount of data to take timely and accurate decisions. There is need of modem tool which not only sift through the vast amount of data but also analyse automatedly and provide rationalized output to human being to interpret it and take informed decisions.There are a lot of inventions and efforts put in using Artificial Intelligence in Cyber Security but it is mainly driven to detect the anomalies or malicious behaviour of IT devices to detect the cyber-attacks / threats in the operations.This approach is not been very successful yet and also these Al models are largely depending on the availability and accuracy of training data which is a key challenge.
[0020] To remediate cyber security threats and vulnerabilities, enterprises need a clear visibility into their attack surface, the critical assets within it, and the threats and risks to those assets. With this visibility, security teams can perform a cyber security vulnerability assessment, prioritize the most critical vulnerabilities, and direct resources to remediate the risks.Therefore, it is essential to discover and map all digital assets to quantify the attack surface.There are tools in the market available for discovery of IT assets, tagging of the assets etc. but their functionality and scope is limited to cover certain types of IT assets, not intelligent enough to determine the interrelation and interdependencies of assets, handle the volume and variety of IT assets in large enterprises and not able to aid IT professionals to take timely and informed decisions accurately.
[0021] Our research and invention are to determine the attack surface by intelligently mapping and ranking IT assets in an enterprise and using the same enterprise’s datasets for training Al model ensuring accuracy and confidentiality of the information.It takes inputs from the tools, security professionals created data and produce the output using Al models to aid security professionals with accurate analysis to take informed decisions on preventive, detective and remedial measures.
[0022] Drawing 2 showcases the block schematic of the Al model used for Methods and Systems for Intelligent Ranking of the IT assets and enhanced prediction of the attack surface.Assets Ranking is the core and heart of the invention.It takes the input of current Assets with labelled data or Tags (201). This data is readily available captured in the existing tools or enterprises build this database over period of time. Data may be tagged with multiple labels but some of the labels (3 labels minimum) which are important to enterprise as per the business model can be give a certain weightage in theMachine Learning algorithm (204) and assets are given ranking based on labels it is tagged for.
[0023] It further takes additional input data which may be captured in the different forms or human created business applications data (202, 203) and another machine learning algorithm (204) corrects tagging of the IT assets and assigns weightages accordingly.This process ensures that any inaccuracy of the labelling of IT assets (201) due to existing tools limitations to capture comprehensive information, localization of the information or non-visibility on the dependencies and usage of IT assets is addressed appropriately in automated way.Number of such inputs to the Al system depends on the comprehensiveness and criticality determined by respective enterprises.
[0024] Al system then ranks the IT assets by final weightage derived for each IT assets and produces output (205) which is in the human readable format and also includes the information about rationale for the final weightages.This ranking help to determine the importance of each IT assets for an enterprise. This ranking list then can be used to determine the threats and risks to enterprise, deploy preventive, detective measures and prioritize remedial actions.
[0025] IT assets ranking output (205) can be used as an input to other Al system which applies input data from vulnerability assessment (206) and security baseline configuration compliance status (207) defined for each type of assets. By making use of Machine Learning algorithm (208), this Al system produces the output (209) in human readable format showcasing vulnerabilities exists against each asset either due to non-patching or missing of baselined security configuration.This showcases the security posture and attack surface in an enterprise. An informed decision can be then taken to reduce the attack surface and prioritize mitigation accordingly.
[0026] Drawing 3 illustrates the hypothesis used for Intelligent Ranking of IT assets.Though this hypothesis is illustrated in relation to IT infrastructure and its exposure to Cyberattacks, this hypothesis can be used for any industrial domain that involves complex variables, variables weigh differently depending on its usage and very difficult for human brains to analyse accurately.
[0027] Artificial intelligence model seeks the input (301) from IT assets database with tags assigned to each asset.Some of the tags that can be considered but not limited are:Operational status - E.g., Operational, non-operational, standby etc.Operational environment - Production, Development, Test, Disaster Recovery etc.Business criticality where assets is used - Category 0, Category 1, Category 2, Category 3 etc.Exposed to internet - Whether asset is Internet Facing, Cloud hosted etc.Accessed from outside of enterprise - Whether Asset is accessed by Internet users, vendors etc.
[0028] Using a Supervised Al model (302), Machine Learning algorithm can assign certain weightages to each selected tag.For example (illustrative purpose only):
[0029] This Al model (303) further seeks input from other data inputs which may be maintained by different application teams or technology teams to document more details about the assets, it’s configuration and dependencies.The Unsupervised model then analyse any missing tags for the assets, analyse its dependencies and interrelations with other IT assets and using machine learning algorithm further adds the weightages if required.For example (illustrative purpose only):
[0030] Al model is further supplied with assets audit log either directly from the IT assets or by a centralized logging system that can help to further determine if it has any further relationship with other IT assets or if assets are accessed from Internet or outside of enterprise etc.The Unsupervised model then analyse any missing tags for the assets, directly accessed from outside of enterprise and interrelations with other IT assets and using machine learning algorithm further adds the weightages if required.For example (illustrative purpose only):
[0031] This Al model (304) then builds hypothesis by combining all weightages from 3 different inputs and ranks asset’s criticality to the enterprise.For example (illustrative purpose only):
[0032] It produces the output report (305) in a human readable format and easy to filter along with the supporting data for the rationale behind this ranking.This report can be then used as an input or basis to various other actions by security professionals. For example, but not limited to:Determine Attack Surface and Indicators of Exposure (Proactive measures)Support or refute cyber-attacks (Detective measures)Vulnerability / Threat Impact assessment (Remedial measures)Individual business application assessmentsThreat modellingSince vulnerabilities and attack vectors are changing continuously and almost daily basis, hypothesis and systems being patented can provide reports and alerts to security professionals in an automated way to take timely and informed decisions.
[0033] Below is end-to-end illustration of one example. a) One IT assets belong to Business Category 2 as tagged in inputl data in
[0028] , so it gets a weightage of 20 b) As received in Input 2 data in
[0029] , this asset is also used by a business Category 0 application (in shared model) hence it gets additional weightage of 10 c) And from Input 3 data in
[0030] , it is determined that this asset is also hosted on cloud and gets additional weightage of 10So, overall weightage for this asset becomes 40 instead of 20 as originally assigned and hence this asset becomes critical for the enterprise as per the ranking table in
[0031]
[0034] Al models used in Hypothesis and Systems make use of combination of Supervised and Unsupervised learning models to resemble some of human characteristics such as:Thoughtfulness: Where it uses past experiences to build the future judgements Self-Learning: With the new information available, it adapts and challenges own understanding and improves decision makingRealtime: Processes information as it receives
[0035] Al systems aid humans to overcome some of the challenges mentioned below but not limited to:Processing overwhelming amount of data gathered every minute and every day Processing of data in different forms from different types of products and technologiesMapping interdependencies and interrelations in the data from different sourcesProcessing speed to take timely decisionsProcessing complex data to take accurate decisions
[0036] Use of Unsupervised Learning method in Al has following high level benefits but not limited to:It can make use of unlabelled dataCapable of finding previously unknown patterns in dataIt is flexible in that it can be applied to a wide variety of problems, including clustering, anomaly detection, and association rule miningIt allows for the exploration of data and the discovery of novel and potentially useful patterns that may not be apparent from the outset
[0037] Use of Supervised Learning method in Al has following high level benefits but not limited to:Supervised learning allows collecting data and produces data output from previous experiencesHelps to optimize performance criteria with the help of experienceIt allows estimating or mapping the result to a new sampleWe have complete control over choosing the number of classes we want in the training data
[0038] Drawing 4 depicts block schematic of the Al System that can be used for predicting enhanced Attack Surface of an enterprise.It makes use of IT Assets ranking output report (401) as one of the inputs. This report ranks the criticality of IT asset to the enterprises based on weightages as below (illustrative purpose only)
[0039] It uses another input from Vulnerability Assessment report (402) which almost all organizations perform with commercial tools which are updated with latest vulnerability database to most current status. This report identifies any unpatched vulnerability exist in all scanned IT assets. Based on criticality of the vulnerability, machine learning algorithm can add weightages to the assets.For example (illustrative purpose only):
[0040] It further uses third input from Security Configuration Compliance Reporting (403) which is carried out with commercially available tools that provides the compliance status of IT assets to respective enterprise defined security configuration baselines. Based on criticality of the noncompliant configuration, machine learning algorithm can add weightages to the assets.For example (illustrative purpose only):
[0041] Al system (404) using supervised learning model can then combine all the weightages against each IT assets of IT ranking output report and builds hypothesis to determine Attack Surface and Threat Exposure of the enterpriseFor example (illustrative purpose only):
[0042] Below is end-to-end illustration of one example. a) One IT assets has criticality ranking of High as per inputl data in
[0023] , so it has weightage of 30 b) In the Vulnerability Assessment Input data in
[0024] , this asset has a open High vulnerability to it, so a weightage of 15 is added to it. c) And from Security Configuration compliance Input in
[0025] , it is non-compliant to a medium baseline configuration. So, a weightage of 10 is added to it.So, overall weightage for this asset becomes 55 and threat exposure to enterprise due to this asset is High.
[0043] It produces the output report (405) in a human readable format and easy to filter along with the supporting data for the rationale behind Threat exposure.This can help security professionals to determine Attack Surface, Threats Exposure and Indicators of Exposures (IOE) to take proactive actions to reduce attack surface of an enterprise.
[0044] Drawing 5 depicts the block schematic of an Al System that can be used for enhanced and more accurate prediction (support or refute) of cyber-attacks.There is huge amount of operational data collected every minute from the IT assets logs.Traditional tools process this data to the possible extent and based on already known normal behaviour patterns, it tries to detect abnormal patterns in the assets or business applications. It has two well-known drawbacks as:Inability to detect new malicious patternMany falls positive indications from the analysisThis is primarily because these tools mostly do pattern or behavioural analysis based on individual assets or sometimes correlation between the logs from related assets determined by security professionals.
[0045] Al system depicted in Drawing 5, takes IT Assets Ranking report (502) as one the inputs that maintains the history of interrelations and interdependencies of the IT assets in the enterprise.Another input (501) is from the tools that detects abnormal behaviour in the IT environment.Al system (503) immediately checks the interdependency or interrelation of the IT asset where abnormality is detected with other IT assets in the network and checks for reporting of any abnormal pattern in those assets that can lead to a cyberattack.This may include but not limited to:Access to the assetData accessEvent timelineCredentials usedIP addresses of sourceAny other parameters
[0046] A weightage to the type of assets may also be provided to related assets that can more accurately determine or refute the cyber-attack.For example (illustrative purpose only}:Abnormality detected on Firewall, Intrusion Detection Systems (IDS), Database Activity Monitoring (DAM) etc. are given higher weightagesAbnormality detected in Routers; Switches may be given other weightages Similarly for Server, Database, Storage and other layered products may be given different weightagesSo, when an abnormality is detected by the tool, Al system can check the pattern in all related assets and if any abnormality is detected (or not detected) then using the weightage of the assets it can conclude or refute the cyber-attack intelligently (504).
[0047] An output report (505) is generated to conclude or refute a cyber-attack with supporting data for security professionals for the further actions.This system may be integrated with alerting tool to send an alert to security professionals in case of cyber-attack conclusion.
[0048] Drawing 6 depicts the block schematic of Al system that determines existing vulnerability or potentially vulnerable assets and generates output report to security professionals to prioritise and take remediation actions.This Al system takes inputs from various sources including but not limited to:Alert notifications from the protection tools (601)Vulnerability Assessment reports (602)Threat Intelligences subscribed by enterprise (603)Alert notifications from IT asset vendors for their respective products
[0049] Machine Learning Algorithm (606) then maps these notifications and vulnerabilities to IT Assets Ranking report (605) which holds the data about all IT assets in the enterprise and ranking its importance to the enterprise.
[0050] It then produces the output report (607) in human readable format that can be further filtered by security professionals to prioritise the assets that need remediation of existing vulnerability or has potential to expose, depending on the ranking of the assets.All inputs are fed in real time and therefore threats to IT assets can be determined on regular basis to keep the enterprise safe from cyber-attacks.
[0051] Drawing 7 depicts the block schematic of Al systems that can be used for Threat Modelling and determining Security Posture of individual business application(s) against cyber-attacks.These are two different Analysis systems but may also be combined together to provide these two functionalities.
[0052] There are many applications used and provided by an enterprise with each application having different nature of business and criticality based on business revenue, regulations, type of data and customers handled etc. to name few. Therefore, it is essential for business heads to know the security posture of businesses they head.Also, due to everchanging landscape of vulnerabilities and ways of exploiting the systems, different threats exist for each application depending on the IT Infrastructure used build the applications. Therefor modelling of threats to application by knowing the IT assets and associated vulnerabilities is very essential now a days.Moreover, this security posture and threat models keep changing constantly in this rapidly changing environment and difficult for security professionals to analyse accurately and in timely manner.
[0053] Al system depicted in Drawing 7 takes IT Assets Ranking report (701) as one of the inputs. This report holds the history of the assets used for a specific business application and their interrelations with other assets used in same business application or other applications in the enterprise.
[0054] It further takes other input from Vulnerability assessment report (703) provided by commercial tool and maps applicable vulnerabilities to the assets.Third input is taken from secure configuration compliance report (706) generated by commercial tool and any non-compliant security configuration is mapped to IT assets.Machine Learning algorithm (702, 707) then analyses the security gaps in individual business applications.
[0055] A report is then generated providing the security posture (704) of individual application(s) for interpretation of security professionals and taking actions thereof.Similarly, Al system can be developed to provide the threat models (705) for individual applications and taking necessary actions by security professionals.The security postures and threats to applications keep changing rapidly and hence these Al systems aid security professionals to keep up with these changes and take informed actions on continuous basis.CLAIMSI claim:1. A method for Artificial Intelligence based apparatus, comprising:• One or more Artificial intelligence models trained with supervised machine learning models, unsupervised machine learning models or combination of both to determine the criticality of IT assets within an enterprise based on combination of but not limited to: i. Usage in the business applications, ii. Operating environment, iii. Exposure to internet, iv. If part of any other critical IT infrastructure, v. Usage by any other critical IT Infrastructure, vi. Access to third party vendors• Forming a hypothesis based on the weightages given to the factors mentioned above and derive the severity of the risk for an enterprise if any of the IT assets gets compromised due to cyber-attacks and intelligently rank all IT assets as per severity.• Generating an output report of this intelligent ranking of the IT assets which can be used by IT professionals in the enterprise for taking informed decision on proactive, detective and remedial measures for those assets.2. The method of claim 1, further comprising:• Artificial Intelligence model trained with machine learning on security gaps as below but not limited to: i. Noncompliance to industry or enterprise defined best practice configuration setting on the IT assets ii. Security vulnerabilities existing in the IT assets• Analysing the data by combining results of above hypothesis with output of Claim 1 for determining Cyber-attack surface for an enterprise.• Generating an output report that predicts enhanced cyber-attack surface for an enterprise and provide Indicators of Exposures (IOE) to IT professional to determine the priority and take informed corrective actions. method of claim 1, further comprising:• Artificial Intelligence model trained with machine learning that seeks alerts from various tools on abnormal behavior in the operations that indicate potential cyber-attack• Analysing any abnormal behavioural pattem(s) in other closely related IT assets as established in claim 1• Forming the Hypothesis based on the weightages given to the IT assets and observing any abnormal behavioural pattern(s) in those IT assets• Generating an output to conclude or refute the cyber-attack method of claim 1, further comprising:• Artificial Intelligence model trained with machine learning that seeks inputs and notifications from various sources as below but not limited to: i. Alerts from cyber protection tools ii. Notifications from respective IT Asset manufacturers iii. Vulnerability Assessments conducted within enterprise iv. Threat Intelligence tools or subscriptions• Applying these alerts or potential security gaps to the Intelligently ranked IT assets in the claim 1• Generating an output report that can be used by IT professionals to determine the criticality and prioritising the remediation measures. method of claim 1, further comprising:• Artificial Intelligence model trained with machine learning that analyses security gaps in the IT assets used directly or indirectly in a specific business application(s) as determined by the hypothesis in claim 1• Determine the security posture of the business application(s) for further actions method of claim 1, further comprising:• Artificial Intelligence models trained with machine learning that analyse security gaps in the IT assets used directly or indirectly in a specific IT Technology platform(s) as determined by the hypothesis in claim 1• Determine the security posture of the technology platform(s) for further actions method of claim 1, further comprising:• Artificial Intelligence model trained with machine learning that analyse IT assets interrelations and interdependencies in a specific business application(s) as determined by the hypothesis in claim 1Derive different Threat Models for the business application(s) An Artificial Intelligence based cyber security apparatus, comprising:• Artificial intelligence model trained with one or more supervised machine learning models, unsupervised machine learning model or combination of both to determine the criticality of IT assets within an enterprise based on combination of but not limited to: i. Usage in the business applications, ii. Operating environment, iii. Exposure to internet, iv. If part of any other critical IT infrastructure, v. Usage by any other critical IT Infrastructure, vi. Access to third party vendors• Forming a hypothesis based on the weightages given to the factors mentioned above and derive the severity of the risk for the enterprise if any of IT assets gets compromised due to cyber-attacks and intelligently rank all IT assets as per severity.• Artificial Intelligence model trained with machine learning on security gaps as below but not limited to: i. Noncompliance to industry or enterprise defined best practice configuration setting on the IT assets based on configuration compliance reports ii. Current security vulnerabilities existing in IT assets based on vulnerability scan reports• Combining the results of above two to generate a report showcasing Cyberattack surface for an enterprise and Indicators of Exposures• Keep this cyber-attack surface information up-to-date based on new vulnerabilities, new security configuration benchmarks or periodic uplifting of security standards by the enterpriseThis output result can be used by IT professionals to prioritise the risk mitigations actions [Preventive actions] to reduce the cyber-attack threats for the enterprise proactively. An Artificial Intelligence based cyber security apparatus of claim 8, further comprising:• Artificial Intelligence model trained with machine learning that: i. Takes the input from third party tool(s) on abnormal pattern in any of the IT assets ii. Check for abnormal pattern in any of the other interrelated, interdependent or handshaking IT assets iii. Forming a hypothesis based on the weightages given to IT assets to conclude the cyber-attack or refute itGenerate an output report with the drilldown data points to support or refute the cyber-attackThis output result can be used by IT professionals to further analyse [Detective actions] the abnormal behaviour for corrective actions.This apparatus can significantly eliminate false positives and provide more accurate predictions on cyber-attacks for IT professionals to act upon. An Artificial Intelligence based cyber security apparatus of claim 8, further comprising:• Artificial Intelligence model trained with machine learning that: i. Takes the input from third party tool(s) such as but not limited to: a. Vulnerability Assessment b. Threat intelligence c. Alerts coming from one or more additional cyber security protection tools d. Notifications from respective IT asset manufacturer ii. Check these alerts or vulnerabilities against intelligently ranked IT assets• Generate an output report listing the IT assets impacted against each alert or vulnerabilityThis output result can be used by IT professionals to further analyse [Remedial actions] the risks, efforts involved and prioritization of remediations.This apparatus will be a significant help for IT professionals as the vulnerabilities and cyber threats keep evolving rapidly and continuously. An Artificial Intelligence based cyber security apparatus of claim 8, further comprising:A mechanism to determine the security postures of individual business application(s) based on the IT assets used in the application and security vulnerabilities and gaps existing in each or any of the IT assets used by the business application.This can give the overview of the cyber security threats and risks existing to the businesses to the business owners, senior management and IT professionals to take informed decisions on preventive, detective and remedial measures. An Artificial Intelligence based cyber security apparatus of claim 8, further comprising:A mechanism to determine the security postures of individual IT technology platform(s) based on security vulnerabilities and security configuration gaps existing in each or any of the IT assets used in the platform.This can give the overview of the cyber security threats and risks existing in the technology platform(s) to take informed decisions on preventive, detective and remedial measures.13. An Artificial Intelligence based cyber security apparatus of claim 8, further comprising:A mechanism to create Threat Models (technical as well as business) based on the IT assets used in the application and security vulnerabilities and security gaps existing in each or any of the IT assets used by the business application.As vulnerabilities and threats are evolving rapidly and continuously, this apparatus will be very useful for simulating new threat models and taking corrective actions.ABSTRACTIt is a method and system which forms a hypothesis that predicts consistent and enhanced cyber-attack surface or exposure in an enterprise.Same type of IT assets is used in different environments, different applications and for different purposes making it very difficult to apply similar criticality or risk category to all IT assets of same type. Multiplying it with different types of IT assets and volume of IT assets in big enterprises makes IT Security professional’s job very complex and cumbersome to make timely decisions.An Artificial Intelligence based system can make use of a data structure containing multiple tags, combine it with cybersecurity vulnerability data and perform analysis to create a consistent and expanding model to provide an intelligent ranking of IT assets for the enterprise and threat exposure for each IT asset. It makes use of multiple machine learning algorithms and combination of supervised as well as unsupervised Al models to produce the results.
Claims
CLAIMS:I claim:
1. A method for Artificial Intelligence based apparatus, comprising:• One or more Artificial intelligence models trained with supervised machine learning models, unsupervised machine learning models or combination of both to determine the criticality of IT assets within an enterprise based on combination of but not limited to: i. Usage in the business applications, ii. Operating environment, iii. Exposure to internet, iv. If part of any other critical IT infrastructure, v. Usage by any other critical IT Infrastructure, vi. Access to third party vendors• Forming a hypothesis based on the weightages given to the factors mentioned above and derive the severity of the risk for an enterprise if any of the IT assets gets compromised due to cyber-attacks and intelligently rank all IT assets as per severity.• Generating an output report of this intelligent ranking of the IT assets which can be used by IT professionals in the enterprise for taking informed decision on proactive, detective and remedial measures for those assets.
2. The method of claim 1, further comprising:• Artificial Intelligence model trained with machine learning on security gaps as below but not limited to: i. Noncompliance to industry or enterprise defined best practice configuration setting on the IT assets ii. Security vulnerabilities existing in the IT assets• Analysing the data by combining results of above hypothesis with output of Claim 1 for determining Cyber-attack surface for an enterprise.• Generating an output report that predicts enhanced cyber-attack surface for an enterprise and provide Indicators of Exposures (IOE) to IT professional to determine the priority and take informed corrective actions.
3. The method of claim 1, further comprising:• Artificial Intelligence model trained with machine learning that seeks alerts from various tools on abnormal behavior in the operations that indicate potential cyber-attack• Analysing any abnormal behavioural pattem(s) in other closely related IT assets as established in claim 1• Forming the Hypothesis based on the weightages given to the IT assets and observing any abnormal behavioural pattern(s) in those IT assets• Generating an output to conclude or refute the cyber-attack4. The method of claim 1, further comprising:• Artificial Intelligence model trained with machine learning that seeks inputs and notifications from various sources as below but not limited to: i. Alerts from cyber protection tools ii. Notifications from respective IT Asset manufacturers iii. Vulnerability Assessments conducted within enterprise iv. Threat Intelligence tools or subscriptions• Applying these alerts or potential security gaps to the Intelligently ranked IT assets in the claim 1• Generating an output report that can be used by IT professionals to determine the criticality and prioritising the remediation measures.
5. The method of claim 1, further comprising:• Artificial Intelligence model trained with machine learning that analyses security gaps in the IT assets used directly or indirectly in a specific business application(s) as determined by the hypothesis in claim 1• Determine the security posture of the business application(s) for further actions6. The method of claim 1, further comprising:• Artificial Intelligence models trained with machine learning that analyse security gaps in the IT assets used directly or indirectly in a specific IT Technology platform(s) as determined by the hypothesis in claim 1• Determine the security posture of the technology platform(s) for further actions7. The method of claim 1, further comprising:• Artificial Intelligence model trained with machine learning that analyse IT assets interrelations and interdependencies in a specific business application(s) as determined by the hypothesis in claim 1• Derive different Threat Models for the business application(s)8. An Artificial Intelligence based cyber security apparatus, comprising:• Artificial intelligence model trained with one or more supervised machine learning models, unsupervised machine learning model or combination of both to determine the criticality of IT assets within an enterprise based on combination of but not limited to: i. Usage in the business applications, ii. Operating environment, iii. Exposure to internet, iv. If part of any other critical IT infrastructure, v. Usage by any other critical IT Infrastructure, vi. Access to third party vendors• Forming a hypothesis based on the weightages given to the factors mentioned above and derive the severity of the risk for the enterprise if any of IT assets gets compromised due to cyber-attacks and intelligently rank all IT assets as per severity.• Artificial Intelligence model trained with machine learning on security gaps as below but not limited to: i. Noncompliance to industry or enterprise defined best practice configuration setting on the IT assets based on configuration compliance reports ii. Current security vulnerabilities existing in IT assets based on vulnerability scan reports• Combining the results of above two to generate a report showcasing Cyberattack surface for an enterprise and Indicators of Exposures• Keep this cyber-attack surface information up-to-date based on new vulnerabilities, new security configuration benchmarks or periodic uplifting of security standards by the enterpriseThis output result can be used by IT professionals to prioritise the risk mitigations actions [Preventive actions] to reduce the cyber-attack threats for the enterprise proactively.
9. An Artificial Intelligence based cyber security apparatus of claim 8, further comprising:• Artificial Intelligence model trained with machine learning that: i. Takes the input from third party tool(s) on abnormal pattern in any of the IT assets ii. Check for abnormal pattern in any of the other interrelated, interdependent or handshaking IT assets iii. Forming a hypothesis based on the weightages given to IT assets to conclude the cyber-attack or refute it• Generate an output report with the drilldown data points to support or refute the cyber-attackThis output result can be used by IT professionals to further analyse [Detective actions] the abnormal behaviour for corrective actions.This apparatus can significantly eliminate false positives and provide more accurate predictions on cyber-attacks for IT professionals to act upon.
10. An Artificial Intelligence based cyber security apparatus of claim 8, further comprising:• Artificial Intelligence model trained with machine learning that: i. Takes the input from third party tool(s) such as but not limited to: a. Vulnerability Assessment b. Threat intelligence c. Alerts coming from one or more additional cyber security protection tools d. Notifications from respective IT asset manufacturer ii. Check these alerts or vulnerabilities against intelligently ranked IT assets• Generate an output report listing the IT assets impacted against each alert or vulnerabilityThis output result can be used by IT professionals to further analyse [Remedial actions] the risks, efforts involved and prioritization of remediations.This apparatus will be a significant help for IT professionals as the vulnerabilities and cyber threats keep evolving rapidly and continuously.
11. An Artificial Intelligence based cyber security apparatus of claim 8, further comprising:A mechanism to determine the security postures of individual business application(s) based on the IT assets used in the application and security vulnerabilities and gaps existing in each or any of the IT assets used by the business application.This can give the overview of the cyber security threats and risks existing to the businesses to the business owners, senior management and IT professionals to take informed decisions on preventive, detective and remedial measures.
12. An Artificial Intelligence based cyber security apparatus of claim 8, further comprising:A mechanism to determine the security postures of individual IT technology platform(s) based on security vulnerabilities and security configuration gaps existing in each or any of the IT assets used in the platform.This can give the overview of the cyber security threats and risks existing in the technology platform(s) to take informed decisions on preventive, detective and remedial measures.
13. An Artificial Intelligence based cyber security apparatus of claim 8, further comprising:A mechanism to create Threat Models (technical as well as business) based on the IT assets used in the application and security vulnerabilities and security gaps existing in each or any of the IT assets used by the business application.As vulnerabilities and threats are evolving rapidly and continuously, this apparatus will be very useful for simulating new threat models and taking corrective actions.
Citation Information
Patent Citations
Systems and methods for network risk management, cyber risk management, security ratings, and evaluation systems and methods of the same
US20220400135A1
Asset risk profiling for enterprise resources
US20230031994A1
Cited By
Internet data security protection system based on artificial intelligence algorithm
CN120128397A
Data security tracing method and system based on artificial intelligence
CN120185949A
Threat intelligence reports generation using artificial intelligence techniques
WO2026206358A1
