Creating a JSON Conduct Policy

Another method of defining a Conduct policy besides creating a policy from the Conduct UI is by uploading a policy file that allows a user to create a very detailed, conditional and fine-grained Conduct Policy. You can create a Conduct policy based on your organizations compliance requirement and upload the policy in Enterprise Archive to filter communications. A Conduct policy must be a .JSON file or a .txt file.

IMPORTANT

Working on Conduct policy requires basic understanding of JSON structure. For basics on creating a JSON document, refer https://json-schema.org/

A policy is a JSON object containing the following two primary fields:

WORDSANDPHRASES

Defines various terms or phrases required by the policy.

To perform a full text search. The search criteria related to text search are specified in this field and is used to search content (Subject, Body, and Attachments) of the indexed documents.

FILTERS

Defines the filter conditions on the corpus of documents to be searched.

To filter documents on various attributes (Participants, Network, Channel, Sent Time and so on) of indexed documents in Enterprise Archive. For example, filtering certain participants of an email communication, documents with a file attachment size and so on.

A Conduct policy contains:

  • One or two sections such as WORDSANDPHRASES and/or FILTERS. The following are four ways in which Conduct policy syntax starts:

    Syntax Representation 1

    Syntax Representation 2

    Syntax Representation 3

    Syntax Representation 4

    {

    "WORDSANDPHRASES": {}

    }

    {

    "FILTERS": {}

    }

    {

    "WORDSANDPHRASES": {},

    "FILTERS": {}

    }

    {

    "FILTERS": {},

    "WORSDANDPHRASES": {}

    }

    The open-close braces {} in the above representations are placeholders to a define search criteria.

  • An entry comprising terms and/or a list of entries are stored in the List Library. For more information, see Configuring List Library.

  • A term which is a word and/or phrase can be combined with proximity operators and/or Boolean operators.

  • Words in combination with wildcard characters enables a single word to match multiple variations of the word.

Also see, Conduct Policy FAQs.

A Conduct policy can be defined as follows:

Standard Policy

A Standard policy uses both WORDSANDPHRASES and FILTERS. See Example - Standard Policy

Lexicon-only Policy

A Lexicon-only policy has TERMS defined either as a LEXICON or LEXICON_LIST without FILTERS. See Example - Lexicon-only Policy

Filter-only Policy

A Filter-only policy has only FILTER filed without any references to LEXICON or LEXICON_LIST. See Example - Filter-only Policy

Important

You must promote a Cognition Studio artifact before it will appear in Enterprise Conduct. Temporary, unpromoted versions won't be available during Enterprise Conduct policy creation. That means your artifact must at minimum have an initial Major, Minor, or Patch version.

Apart from the WORDSANDPHRASES or FILTERS, the POLICY_TYPE field is allowed as a top level field. This is a filter condition based on the policy types in a Conduct queue. Policy_Type can have the following values:

FLAG

Used to fetch communications that are required to go through the Conduct workflow. This is the default policy type.

IGNORE

Used to skip or ignore communications that are not required to go through the Conduct workflow. IGNORE is used in tandem with one or more policy of Flag type and leads to a kill-list of documents which are filtered out from list of documents fetched by Flag policies. Considering the limited use of Ignore policies, document ID is the only detail of documents which is fetched.