Assessing and managing computational risk involved with integrating third party computing functionality within a computing system
In general, various aspects of the present disclosure provide methods, apparatuses, systems, computing devices, computing entities, and/or the like for addressing a modified risk rating identifying a risk to an entity of having computer-implemented functionality provided by a vendor integrated with a computing system of the entity. In accordance various aspects, a method is provided that comprises: receiving a first assessment dataset for computer-implemented functionality; detecting an inconsistency between a value of an attribute for the computer-implemented functionality specified in the first assessment dataset and a corresponding value of the attribute specified in a second assessment dataset for the computer-implemented functionality; modifying a risk rating that identifies a risk to the entity of having the computer-implemented functionality integrated with the computing system to generate a modified risk rating based on the inconsistency; and in response, performing an action with respect to the computing system to address the modified risk rating.
1. A method comprising:
providing, by computing hardware, a first question from a set of questions found in an electronic assessment for display on a user interface, wherein the user interface solicits a first answer to the first question, and the set of questions relates to computer-implemented functionality provided by a vendor;
receiving, by the computing hardware and via the user interface, the first answer to the first question to provide a first question/answer pairing;
mapping, by the computing hardware, the first question/answer pairing to an attribute related to the computer-implemented functionality;
mapping, by the computing hardware, the attribute to a second question/answer pairing, wherein the second question/answer pairing:
is related to the attribute,
comprises a second question of the set of questions, and
comprises a second answer provided to the second question;
comparing, by the computing hardware, the first answer to the second answer to identify an inconsistency in the first answer with respect to the second answer;
determining, by the computing hardware, to address the inconsistency; and
responsive to determining to address the inconsistency, performing, by the computing hardware, an action to address the inconsistency, wherein the action comprises at least one of:
(1) prompting, via the user interface, to provide at least one of supporting information or supporting documentation to address the inconsistency, (2) providing, via the user interface, a follow up question related to the inconsistency, or (3) requesting, via the user interface, the inconsistency in the first answer to be redressed.
2. The method of claim 1 , wherein determining to address the inconsistency comprises processing the inconsistency via a decision engine to generate a relevance of the inconsistency, the decision engine comprising a rules-based model configured to use a set of rules in determining a level of the inconsistency as a measure of the relevance of the inconsistency.
3. The method of claim 1 further comprising determining, by the computing hardware, the action to perform by:
processing at least one of the first question/answer pairing or the second question/answer pairing using a machine-learning model to generate a data representation having a set of predictions in which each prediction is associated with a particular action to take to address the inconsistency; and
selecting, based on the set of predictions, the action to address the inconsistency.
4. The method of claim 3 , wherein the machine-learning model is trained using training data derived from responses to follow up requests previously provided for inconsistencies detected in past assessment question/answer pairings that comprises at least one of (1) whether the inconsistencies detected in the past assessment question/answer pairings were ignored, (2) whether related follow up actions for the inconsistencies detected in the past assessment question/answer pairings were ignored, (3) particular types of actions taken in response to the inconsistencies detected in the past assessment question/answer pairings, or (4) at least one of a framework or standard that is mapped to the past assessment question/answer pairings.
5. The method of claim 1 , wherein mapping the first question/answer pairing to the attribute related to the computingcomputer-implemented functionality and mapping the attribute to the second question/answer pairing is performed via a data structure that maps the first question/answer pairing and the second question/answer pairing to the attribute.
6. The method of claim 1 , wherein the first answer and the second answer are in a freeform text format and comparing the first answer to the second answer to identify the inconsistency in the first answer with respect to the second answer involves:
performing a natural language processing technique on the first answer to generate a first embedded representation of the first answer,
performing the natural language processing technique on the second answer to generate a second embedded representation of the second answer, and
comparing the first embedded representation to the second embedded representation to identify the inconsistency.
7. The method of claim 1 further comprising determining, by the computing hardware, the action to take to address the inconsistency based on at least one of a type of the attribute or a past response to a past inconsistency related to the attribute.
8. A system comprising:
a non-transitory computer-readable medium storing instructions; and
a processing device communicatively coupled to the non-transitory computer-readable medium,
wherein, the processing device is configured to execute the instructions and thereby perform operations comprising:
providing a first question found in an electronic assessment for display on a user interface, wherein the user interface solicits a first answer to the first question, and the first question relates to computer-implemented functionality;
receiving, via the user interface, the first answer to the first question to provide a first question/answer pairing;
mapping the first question/answer pairing to a plurality of attributes related to the computer-implemented functionality;
mapping the plurality of attributes to a second question/answer pairing, wherein the second question/answer pairing comprises a second question, and a second answer provided to the second question;
comparing the first answer to the second answer to identify an inconsistency in the first answer with respect to the second answer;
determining to address the inconsistency; and
responsive to determining to address the inconsistency, performing an action to address the inconsistency, wherein the action comprises at least one of: (1) prompting, via the user interface, to provide at least one of supporting information or supporting documentation to address the inconsistency, (2) providing, via the user interface, a follow up question related to the inconsistency, or (3) requesting, via the user interface, the inconsistency in the first answer to be redressed.
9. The system of claim 8 , wherein determining to address the inconsistency comprises processing the inconsistency using a rules-based model configured to use a set of rules in determining a level of the inconsistency as a measure of a relevance of the inconsistency.
10. The system of claim 8 , wherein the operations further comprise determining the action to perform by:
processing at least one of the first question/answer pairing or the second question/answer pairing using a machine-learning model to generate a data representation having a set of predictions in which each prediction is associated with a particular action to take to address the inconsistency; and
selecting, based on the set of predictions, the action to address the inconsistency.
11. The system of claim 10 , wherein the machine-learning model is trained using training data derived from responses to follow up requests previously provided for inconsistencies detected in past assessment question/answer pairings that comprises at least one of (1) whether the inconsistencies detected in the past assessment question/answer pairings were ignored, (2) whether related follow up actions for the inconsistencies detected in the past assessment question/answer pairings were ignored, (3) particular types of actions taken in response to the inconsistencies detected in the past assessment question/answer pairings, or (4) at least one of a framework or standard that is mapped to the past assessment question/answer pairings.
12. The system of claim 8 , wherein mapping the first question/answer pairing to the plurality of attributes related to the computer-implemented functionality and mapping the plurality of attributes to the second question/answer pairing is performed via a data structure that maps the first question/answer pairing and the second question/answer pairing to the plurality of attributes.
13. The system of claim 8 , wherein the first answer and the second answer are in a freeform text format and comparing the first answer to the second answer to identify the inconsistency in the first answer with respect to the second answer involves:
performing a natural language processing technique on the first answer to generate a first embedded representation of the first answer,
performing the natural language processing technique on the second answer to generate a second embedded representation of the second answer, and
comparing the first embedded representation to the second embedded representation to identify the inconsistency.
14. The system of claim 8 , wherein the operations further comprise determining the action to take to address the inconsistency based on at least one of (1) a number of the plurality of attributes, (2) a type of each of the plurality of attributes, or (3) one or more past responses to one or more past inconsistencies related to the plurality of attributes.
15. A non-transitory computer-readable medium having program code that is stored thereon, the program code executable by one or more processing devices for performing operations comprising:
providing a first question found in an electronic assessment for display on a user interface, wherein the user interface solicits a first answer to the first question, and the first question relates to computer-implemented functionality;
receiving, via the user interface, the first answer to the first question to provide a first question/answer pairing;
mapping the first question/answer pairing to an attribute related to the computer-implemented functionality;
mapping the attribute to a second question/answer pairing, wherein the second question/answer pairing comprises a second question, and a second answer provided to the second question;
comparing the first answer to the second answer to identify an inconsistency in the first answer with respect to the second answer;
determining to address the inconsistency; and
responsive to determining to address the inconsistency, performing an action to address the inconsistency, wherein the action comprises at least one of: (1) prompting, via the user interface, to provide at least one of supporting information or supporting documentation to address the inconsistency, (2) providing, via the user interface, a follow up question related to the inconsistency, or (3) requesting, via the user interface, the inconsistency in the first answer to be redressed.
16. The non-transitory computer-readable medium of claim 15 , wherein determining to address the inconsistency comprises processing the inconsistency using a rules-based model configured to use a set of rules in determining a level of the inconsistency as a measure of a relevance of the inconsistency.
17. The non-transitory computer-readable medium of claim 15 , wherein the operations further comprise determining the action to perform by:
processing at least one of the first question/answer pairing or the second question/answer pairing using a machine-learning model to generate a data representation having a set of predictions in which each prediction is associated with a particular action to take to address the inconsistency; and
selecting, based on the set of predictions, the action to address the inconsistency.
18. The non-transitory computer-readable medium of claim 15 , wherein mapping the first question/answer pairing to the attribute related to the computer-implemented functionality and mapping the attribute to the second question/answer pairing is performed via a data structure that maps the first question/answer pairing and the second question/answer pairing to the attribute.
19. The non-transitory computer-readable medium of claim 15 , wherein the first answer and the second answer are in a freeform text format and comparing the first answer to the second answer to identify the inconsistency in the first answer with respect to the second answer involves:
performing a natural language processing technique on the first answer to generate a first embedded representation of the first answer,
performing the natural language processing technique on the second answer to generate a second embedded representation of the second answer, and
comparing the first embedded representation to the second embedded representation to identify the inconsistency.
20. The non-transitory computer-readable medium of claim 15 , wherein the operations further comprise determining the action to take to address the inconsistency based on at least one of a type of the attribute or a past response to a past inconsistency related to the attribute.