Run investigation

Once Adverse Media Monitoring is installed and the configuration is complete, you can initiate investigations in one of the following ways:

• Start the Core or Ad Hoc Business Process from Control Tower
• Use the Ad Hoc investigation search tool
  • Start Ad Hoc investigation
  • Run Ad Hoc search
• Call AMM REST API
• Ingest articles and entity details via a CSV file

Start Business Process from Control Tower

You can initiate Adverse Media Monitoring in Control Tower by running the Core or Ad Hoc Investigation Business Processes (BP). The main difference is that the Ad Hoc Investigation BP creates an Ad Hoc search screen that enables analysts and business users without access to Control Tower to initiate investigations from Workspace, for instance, as part of a transaction monitoring fraud alert investigation.

To start the Core BP, do the following:

1. In the Control Tower main menu, click AI Agents > View All and find the Evan: Adverse Media Monitoring group.

   

2. In the group, find the AMM variation you wish to use and click it.

3. On the variation's details page, find the Adverse Media Monitoring BP and click it to open its process definition.

   

4. Next to a core BP instance, click the Actions menu button (...) and choose one of the following options:

   • If the instance is not running, click Edit and Run. The Edit Business Process window appears.

   • To create another instance of the process, click Copy. Thus, you can run a new instance without impacting an existing one.

   

5. In the Business Process window, go to the Data tab and click Upload Data.

   

6. In the Upload Your Main Data window, drag and drop a CSV input file and click Upload. The data is uploaded to the BP.

   note

   If you are using the Human in the Loop feature, you can pre-assign tasks to workers in your organization. To do this, enter the worker’s ID in the _sys_mt_hit_assignee column for the required entity rows of the input CSV file.

   Currently, real-time validations are not supported. If a BP can't recognize a value in the sys_mt_hit_assignee column, the associated task is not assigned to any worker.

7. Go to the Run tab and click Run this process to start the BP.

If the Manual Task is enabled, the investigations appear in Workspace as they are completed. If it is disabled, the reports are generated as configured.

note

Occasionally, no results are found for the search. If this occurs, a report for each entity is created. A separate email with a link to view and download the report is sent for each entity that does not have search results. A copy of this report is also stored in the location specified during the AMM configuration.

Once all the entities within the input file have been processed, the Business Process ends automatically.

Run via Ad Hoc investigation search tool

The Ad Hoc investigation is a Business Process that allows customers to trigger Adverse Media Monitoring from the Workspace user interface. It is available after installing and configuring Adverse Media Monitoring.

Start Ad Hoc investigation

To start an Ad Hoc investigation, do the following:

1. In the Control Tower menu, click AI Agents > View All and find the Evan: Adverse Media Monitoring group.

   

2. In the group, find the AMM variation you wish to use and click it.

3. On the variation's details page, find the Adverse Media Monitoring Ad Hoc investigation BP and click it to open its definition.

   

4. Next to a BP instance, click the Actions button (...) and select one of the following options:

   • If the instance is not running, click Edit and Run. The Edit Business Process window appears.

   • To create another BP instance, click Copy. Thus, you can run a new instance without impacting an existing one.

   

5. In the Business Process window, go to the Run tab and click Run This Process to start the Business Process.

info

The number of Ad Hoc search tasks created in Workspace is configured in the Investigation step.

Run Ad Hoc search

To run an investigation using the Ad Hoc search screen, do as follows:

1. Log in to Workspace, find Submit Adverse Media Monitoring Investigation request vX.X, and open it.

   

   If you do not need all Entity information fields, you can leave only required fields during the user interface configuration on the Investigation step: Ad-hoc investigation settings > Fields to include in the ad-hoc investigation task. For details, see Install and configure | Investigation.

   

2. In the Entity information group, enter data for the entity you wish to screen. Pay attention to the following considerations:

   • Only Entity name is required to initiate a search. All other fields are optional but encouraged since more information on the entity enables the automation to eliminate more false positive results.

   • To allow re-applying decisions in future investigations for the same entity, provide a custom Entity ID. The same Entity ID must be used for all further investigations on the entity.

   • You can either upload data in a CSV file or enter the data manually.

3. Review the settings in the Requests details group.

   • If necessary, in the Search Time Period box, select a different time for scanning.

   • Other fields are configured in Control Tower and listed on this screen for reference only.

4. Review the settings in the Results group:

   Enter one or more email addresses where to send the results. In the Send copy to field, it is recommended to specify a group email so that other team members can refer to previous investigations and their results.

5. Click the Submit button.

Once the results are ready, an email notification is sent with a link to the assignments in Workspace and a final report.

If your organization does not use the Human in the loop feature, you only receive a final report. This is also indicated on the Ad Hoc search screen.

note

In some cases, no results are found for the search. If this occurs, a report is created for each entity that does not have search results. A separate email with a link to view and download the report is sent for each such entity. A copy of this report is also stored in the S3 storage specified during the AMM configuration.

Call AMM REST API

Adverse Media Monitoring can be invoked by calling the WorkFusion AMM REST API. To work with API, you must first configure a variation as it requires a BP associated with it. For details, see the Install and configure | REST API endpoints guide.

For more information on REST API, read Running AI Agents via REST API.

Ingest via CSV file

To run an investigation using an uploaded file, follow the steps below:

1. Configure ingestion-related AMM settings.

2. Prepare and upload the input file for the AMM to ingest.

3. Run the Adverse Media Monitoring File Ingestion Business Process.

Configure file ingestion

To set Evan to ingest data from a CSV file, follow the steps below:

1. In the Control Tower menu, click AI Agents > View all and find the Evan: Adverse Media Monitoring group.

   

2. Find the AMM variation you wish to use and click it.

3. On the variation's details page, open the Actions menu for the variation and select Settings to open the configuration screen.

   

4. On the Input step, in the Search provider dropdown box, make sure to select External Source.

   

5. Expand the External Source section and configure the input settings as required for the provider. For detailed instructions, read Install and configure | External Source.

6. If you want to use metadata information to disposition articles first, go to the Investigation step and select Yes to answer the Do you want to use available meta data to adjudicate articles first? question.

   

Upload input

To prepare input for ingestion, follow the steps below:

1. Go to MinIO (sample MinIO URL: https://amm-release-minio-ui.cloud.workfusion.com/) and log in with SSO.

2. Find the bucket you specified when configuring the external source.

3. Navigate to the directory (or create one if it doesn’t exist) you specified when configuring the external source provider.

4. Click Upload > Upload File.

   

5. Select an input CSV file and upload it.

   To make sure your CSV file has the correct format, download the sample and check your file against it. Files that do not comply with the format requirements are not ingested.

After some time, the file is picked up by the AI Agent and moved to the processed folder.

Run File Ingestion Business Process

To start the Adverse Media Monitoring File Ingestion BP, follow the steps below:

1. In the Control Tower menu, click AI Agents > View all and find the Evan: Adverse Media Monitoring group.

   

2. Find the Adverse Media Monitoring variation you wish to use and click it.

3. On the variation's details page, click the Adverse Media Monitoring File Ingestion BP to expand the process definition.

   

4. Choose a BP instance, open the Actions menu for it, and select Edit and Run.

   

5. If you configured AMM to use metadata information to disposition articles first, go to Workflow tab, click the Execute NSS step, and specify the Signal ID from the NSS Business Process. Click Save to save the changes.

   

6. Navigate to the Run tab and click Run This Process.

Review CSV fields

In an input CSV file, rows are grouped by uuid:

• Every row with the same uuid describes one entity, and each of those rows contributes one article to that entity.

• Entity-level attributes are read from the first row of the group. Article-level fields are read from every row.

For field-level validation rules (country ISO codes, date formats, year ranges, length limits), see Learn about input and output data | Input data validation.

Required columns

Column	Type	Description
uuid	String	Grouping key for the entity. Rows sharing a uuid form one entity plus its articles. Must be present on every row.
search_request	String	Entity name or search term. Max 256 characters.
provider	String	Provider that processes the entity or supplied the article.

Entity identity and attributes

Column	Type	Description	Legacy alias
entity_id	String	Client-side entity identifier; max 256 characters	—
entity_type	Enum	Individual or Company (case-insensitive)	—
gender	Enum	Male or Female (case-insensitive)	—
year_of_birth_incorporation	Integer	Year of birth (individual) or incorporation (company)	year_of_birth, yob
date_of_birth_incorporation	String (date)	Date of birth / incorporation	date_of_birth
residence_operating_country	String (ISO-2)	Residence (individual) or operating (company) country	residence_country
country	String (ISO-2)	Country; used as the final fallback for residence / operating country	—
citizenship_incorporation_country	String (ISO-2)	Citizenship (individual) or country of incorporation (company)	citizenship_country
state	String	State	—
subdivision	String	Subdivision/state ISO-3166 code, country-specific alpha code, or unique subdivision name	—
city	String	City. Requires a residence/operating country when city validation is enabled.	—
entity_additional_information	String	Free-text notes about the entity	—

Scheduling and workflow metadata

Column	Type	Description
batch_id	String	Batch identifier.
batch_size	Integer	Batch size.
cost_center_code	String	Cost center code.
cost_center_name	String	Cost center name.
sla_date	String (ISO-8601 datetime)	Service Level Agreement (SLA) due date, yyyy-MM-dd'T'HH:mm:ss.
sla_duration	String	SLA duration.
business_line	String	Business line.
screening_service	String	Screening service.
investigation_date	String (ISO-8601 datetime)	Investigation date, yyyy-MM-dd'T'HH:mm:ss.
initiator	String	Investigation initiator.
related_entity_id	String	Related entity identifier.
add_to_ongoing_monitor	Boolean	Enroll the entity in ongoing monitoring. Parsed leniently: only true is true; blank or other is false.
ongoing_monitoring	Boolean	Ongoing monitoring flag (true/false); blank leaves it unset.
investigation_uuid	UUID	Original investigation UUID. Must be a valid UUID string.
review_availability_date	String (ISO-8601 datetime)	When the review becomes available, yyyy-MM-dd'T'HH:mm:ss.
assignee	String	Human-in-the-loop pre-assignment (worker ID) for the entity's task.
assigned_user_groups	String	Pre-assigned user groups.
priority	Integer	Task priority.
search_period	Integer	Generic search period (days). Bound for backward compatibility; not present in the template. Prefer the provider-specific days_search_period_* columns.

Provider search periods

Each restricts the look-back window (in days) for one provider. All values are optional and have the integer format.

Column	Applies to
days_search_period_google	Google
days_search_period_brave	Brave
days_search_period_factiva_headlines	Factiva Headlines
days_search_period_lexis_nexis_lp_media	LexisNexis L&P Media
days_search_period_world_check_one	World-Check One
days_search_period_external_source	External Source

Article fields

One article is built per row (within the uuid group). A row produces an article only if at least one of article_url, title, summary, content, content_link, or article_additional_information is non-blank. Otherwise, the row contributes no article.

Column	Type	Description
article_url	String (URL)	Link to the article.
title	String	Article title.
summary	String	Article summary.
content	String	Inline article content (HTML or plain text).
content_link	String (URL)	Link to where the article content is stored.
article_additional_information	String	Free-text notes about the article.
investigation_status	Enum	Pre-disposition status.
external_review_comment	String	Reviewer comment carried with the article.
external_provider_search_url	String (URL)	Link to the external provider's search result.
publication_date	String (date)	Article publication date.
relevance_score	String (number)	Article relevance value.

Provider-specific guidance

• External Source: the provider consumes the following article fields from the input file: article_url, title, summary, content, content_link, article_additional_information, investigation_status, and external_review_comment.

  Below is a sample CSV input for an entity with an article URL (External Source):

  uuid,search_request,provider,entity_type,gender,year_of_birth_incorporation,article_url,title,summary,investigation_status,external_review_comment,publication_date
  1234,Amanda Riley,External Source,Individual,Female,1986,https://www.example.com/story,Woman faked cancer to swindle donors,Article details an individual who faked a cancer diagnosis to collect donations.,Needs Investigation,Confirmed adverse media match,2023-05-30

• Other providers (Google, Brave, Factiva, LexisNexis L&P, World-Check One)

  Internal search providers use the entity attributes and the matching days_search_period_* column. Article-link fields, such as article_url, title, summary, content, content_link, are not used by internal search. Supply them only for External Source review.

  Below are sample CSV inputs:

  • Minimal entity, no URL

    Only the three required columns are populated. Everything else is left blank.

    uuid,search_request,provider
    ab123,Anna Banana,Google

  • Entity with inline title, summary, content and no URL

    uuid,search_request,provider,entity_type,country,city,title,summary,content,article_additional_information,investigation_status
    c789,Global Trading LLC,External Source,Company,GB,London,Regulator fines Global Trading LLC,Summary of a regulatory penalty issued against the company for sanctions breaches.,The regulator announced a settlement following an investigation into export-control violations.,inline article supplied without a URL,Needs Investigation