Skip to main content

Reveal Review Publication

Reveal AI Terminology

Term

Description

Reveal AI Handling

Theme

Story

The original vision of Reveal AI was Data will tell us stories. Counsel construct a narrative of a case, whose story is found in communications and documents generated in the course of the dispute.

Reveal AI analyzes the use of language for sentiment, emotion and theme as well as the words in the documents to not only identify but illustrate relevance.

Storybook

The documentary data of a case. A collection of content and work product equivalent to a case, project or workspace in other review platforms.

Think of a storybook as a compilation of characters, actions, motivations and conversations underlying a conflict. Reveal AI applies machine learning and analytics to solving the mystery of how the conflict came to be, who was involved, and inferring their emotions along the way.

Segment

A segment is an individual email; every segment has only one writer; without any replies or forwards.

As an example, I write about Acme Corp one time and three people receive this email. This email is considered one segment. If someone replies to the email and includes the original email, there is only one new segment added to the Reveal AI project.

Reveal AI counts individual segments to tell the story of how many times something was written about.

Reveal AI uses a content based algorithm to deduplicate segments in the project. This greatly reduces the amount of duplicative data in the project.

Emails

Document

A document is an email that was originally created and may contain one or more segments; a reply to an email has both the new segment and contain the previous segment.

Reveal AI globally deduplicates all documents based on given hash value. Reveal AI maintains the link between segments and documents.

Emails

Copy

A copy is a deduplicated document.

Every document imported is recorded in the Reveal AI database. Only one copy of document is exposed to end-users.

Emails

Thread

A thread is a document or series of documents that shares at least one exact segment.

For example, an email may be sent to a recipient, who in turn forwards to another recipient and starts a parallel conversation.

Reveal AI detects threads and creates a ThreadID.

When searching within Reveal AI, Reveal AI will always retrieve the full thread when interrogating a hit.

Emails

Inclusive Document

An inclusive email is a document that contains the maximum amount distinct segments for a thread.

As an example, an exchange of emails will create multiple documents. However, the final email will have all segments from the previous communications. An end-user only has to interrogate the final document to read all possible segments.

There may be multiple inclusive documents within a thread. For example, multiple conversation branches may create multiple inclusive documents.

Reveal AI always returns all inclusive documents within a thread.

Emails

Baseball/Profile Card

A baseball card is a profile Reveal AI dynamically builds around any Entity, Summary Phrase or Communicator that an end-user is interrogating.

Reveal AI creates a baseball card to quickly familiarize the end-user with the data being interrogated.

The baseball card provides pertinent high-level statistics and narrative.

Baseball Card

Visualization (Viz)

The visualization graphically displays communications between people or about subjects.

Reveal AI creates a link chart based on the baseball data being interrogated.

Visualization

Time chart

The visualization graphically displays communications as distributed over a selected period of time.

Reveal AI creates an interactive chart based on the chronology of communications.

Visualization

Named Entity

A named entity is an extracted piece of data identified by proper name by the Reveal AI.

Named Entities are an efficient method for searching (filtering) for specific people, places or things.

For example, imagine keyword searching for Washington. A keyword search cannot distinguish between a state, street, person or school. Using a named entity, the search results will be more precise.

Reveal AI analyzes every piece of data in the system and identifies the named entities.

Reveal AI identifies over 21 entity types and more are possible with training.

Person

(discussion about a person)

603e748e7876c.png

Entities

Geo-political

(city, state, country)

603e7490096a8.png

Organization (company)

603e74914d446.png

Money

(currency discussed)

603e74928961a.png

Temporal

(dates discussed)

603e7493c34e8.png

Law

(legal jargon)

603e74950cf50.png

Quantity (measurements)

603e749631b9c.png

Groups

(e.g. Democrats, Republicans)

603e74976ebc3.png

Location

(specific areas)

603e74914d446.png

Technology (jargon)

603e749a45fbb.png

Topic

(conceptual focus of sentence)

603e749bccef6.png

Category

(concept contained in document; hierarchical to Super Category)

603e749d48e21.png

Event

(e.g. Super Bowl)

603e749ea518f.png

Ordinal

(numbers)

603e749631b9c.png

Percent

603e74a07f1fb.png

Product

(discussion about a product)

603e74a1da5aa.png

Work of Art

(discussion about music, books, etc.)

603e74a382842.png

Summary Phrase

(extracted important phrases from documents)

603e74a50802f.png

Law Firm

(Law Firms mentioned in document)

603e74a6460ef.png

Super Category

(broad concept contained in document; hierarchical to Category)

603e74a768e08.png

Custom Entity

A Custom Entity is an entity defined by a piece of data and identified by proper name by the Reveal AI.

Named Entities are an efficient method for searching (filtering) for specific people, places or things.

For example, imagine keyword searching for Washington. A keyword search cannot distinguish between a state, street, person or school. Using a named entity, the search results will be more precise.

Reveal AI analyzes every piece of data in the system and identifies the named entities.

Reveal AI allows custom entities to be incorporated into Models which can be applied to characterize project data.

Entities

Entity Model

PER, ORG, Phone Number, Drug Name

Reveal AI uses defined entities to train models in order to organize and characterize project data.

Entities

Entity Type

The category of entity, e.g., Person, Organization, Phone number, Group, Product or Topic

Reveal AI uses combinations of these entity types in building custom models.

Entities

Model Library

  • Entity Model Library

  • COSMIC Model Library

In addition to the Standard Model, models created in storybooks may be published to the model library of the storybook or for all of a tenant’s storybooks to be reused, adapted and re-saved.

Models

Term Report

Search tool used in complex querying and reporting. Creates a saved search; also used in custom entity creation.

Reveal AI uses lists of keyword terms, entities, Boolean and proximity or regular expressions to create reports and, in the case of Entity search and extraction, uses hits to create custom entity types and models.

Entities

Model

A portable package of knowledge built from custom entity types.

For example, “obscenity” and “suggestion” examples would contribute to a custom entity type of “harassment” as the basis of a model.

Used by Reveal AI to apply a set of categorization rules in gathering and analyzing classes of information; model examples set a mathematical start and end point for relevance.

Models

Summary Phrase

A summary phrase is a keyword or phrase which is highly important to segment.

Summary phrases can summarize any collection of segments to their most important word or phrases.

Reveal AI uses a sophisticated statistical algorithm to rank keyword and phrases in segments.

An example may be the keyword of Sincerely. Sincerely or Thanks may be present in a document and also located in the majority of the email segments. These words would receive a low ranking compared to a word like fire which may be used infrequently but is extremely within a collection.

Entities

Topic

Topics are the most representative and informative phrases we can extract from text.

Reveal AI computes Topics based on the summary phrases from each segment and also from the context of the conversation in the whole thread.

At segment level, a document could have multiple Topics.

Entities

Launchpad

The Launchpad panel to the left of the screen is where faceted search and entity listings may be selected to open details and visualizations for the item selected.

The Reveal AI LaunchPad lists global and end-user information about project.

Layout

Sentiment

Sentiment is the tone of a segment. The ability to filter on segments with a specific tone can add context to the story of a project.

Reveal AI uses an algorithm to calculate a score of the sentiment for the overall segment.

For example, a writer may say:

Everything is a mess and we need to shut this down right away. However, the staff is nice.”

The segment contains both negative and positive sentiment. Reveal AI calculates an overall score to take this scenario into account.

Sentiment

Negative Sentiment

Negative sentiment equals a writer using words and tone associated with negative connotation.

A segment with a score lower than -1 will equal a negative segment in Reveal AI.

Sentiment

Positive Sentiment

Positive sentiment equals a writer using words and tone associated with positive connotation.

A segment with a score higher than 1 will equal a positive segment.

Sentiment

Communication

A communication is a pair of senders/receivers that engage in active conversation.

Reveal AI collects and analyzes and presents visualizations of communications weighted for frequency, time, topic, sentiment and other factors.

Communication

Writer

A writer of a segment

Reveal AI identifies the sender of an email as Writer (or Writers).

Communication

Reader

A reader of a segment

Reveal AI Identifies anyone addressed in the To, CC or BCC field as a Reader.

Communication

Spectator

A spectator is a reader who indirectly received a segment.

Reveal AI identifies the readers who were forwarded emails and were not part of the original active conversation.

This can be useful to understand what an actor is learning via forwarded emails.

Communication

Work Shift

A work shift is a bucket of time in a day. Filtering data by the work shift can further give context to the story and intent of the writer.

Reveal AI identifies the following work shifts:

  • Biz Hours: Mon-Fri (7:00am to 6:00pm)

  • Biz Eve: Mon-Thur (6:00pm to 12:00am)

  • Off-Hours: Fri-Sun & Mon-Fri 12:00 am to 7:00am

Dates

Recipient Count

The recipient count is the number of addresses on the TO, CC and BCC lines.

Reveal AI allows end-users to select the maximum recipient count to remove potential “blasts of emails.”

Communication

Global Reciprocal Ratio

Number of emails sent vs. number of emails received. A large recipient ratio often indicates the person associated with the email address might be spam or mass-marketing.

Reveal AI calculates the global reciprocal ratio using number of emails the person sent divided by number of emails the person received.

Communication

“About” Mode

The “About” mode retrieves emails in which the person was mentioned, but he or she is NOT included in the conversation.

The About Feature looks for the situation where a person’s name is mentioned in the body of an email and the person mentioned is not contained in the metadata features (to/from/cc/bcc) of the email. The power of this feature is that the named is resolved to a normalized proper name (Bill ---> William Jones) for easier searching by the End User.

Communication

Patterns

A Pattern is a data anomaly identified with the collection.

An identified pattern isolates a small collection of segments where this behavior breaks the typical day to day activity.

A pattern may find vignettes of time that were greatly stressful to an organization.

Reveal AI identifies unusual behavior based on the following criteria:

  • Date Range

  • Sender or receiver with Specific People or groups of people

  • Concepts within segments

  • Tone

  • Named entities discussed.

  • Domain utilized

  • Minimum frequency of 6 occurrences within a 10-day period.

Pattern

“Self” Mode

The “Self” mode retrieves emails in which the person sent email to himself or herself.

Reveal AI detects emails the person sent to himself or herself after name resolution. This feature can be used to capture emails sent from a person’s working email address to a personal email address.

Communication

Pseudonyms

Alternative names detected for people, e.g., William Jones’s pseudonyms may be: Bill Jones, William B. Jones, etc.

Reveal AI utilizes disambiguation and name resolution technology to collapse entities even if names are misspelled, different email addresses, nicknames or acronyms are used.

Entities

Mention

A mention is an alternative name found for a Named Entity, e.g., New York may have mention of NY.

Reveal AI utilizes disambiguation and name resolution technology to collapse entities even if names are misspelled, different email addresses, nicknames or acronyms are used.

Entities

Category

A Category is a broad concept found in a document that may refer to an industry or object. End-users can utilize Categories to search for granular common name conceptual terms located in the documents.

A Category is comprised of one or more topics found in the documents. For example, the Category accounting may include the topics accounting issues & internal accounting controls.

Reveal AI utilizes Advanced AI to detect concepts that are being discussed without expensive human training.

Reveal AI classifies thousands of categories.

Entities

Super Category

A Super Category is a broad concept found in a document that may refer to an industry or object. End-users can utilize Super Categories to search for common name conceptual terms located in the documents.

A Super Category is comprised of one or more categories related to the concept and found in documents. For example, the Super Category Finance may include the categories accounting & debt settlement.

Reveal AI utilizes Advanced AI to detect concepts that are being discussed without expensive human training.

Reveal AI classifies thousands of categories.

At segment level Reveal AI may find one or more Super Categories present, so that an attachment discussing finance may have a sidebar topic addressing the super category of construction.

Entities

Cluster

Clustering analyzes textual documents and groups conceptually similar documents. Clusters are automatically generated, and display in the Tree Map view.

Clustered documents may relate to a subject or a type of communication. For example, documents in an Earnings Call cluster would be given a higher cluster score for being closer to the center of the cluster (for example, a report or transcript of such a call as opposed to preparatory discussions).

Reveal AI utilizes Advanced AI to group conceptually close documents together.

When clustering at thread level, an attachment could go to a different cluster.

Clustering

Recommend Segment

Utilize Reveal AI to recommend additional segments (documents) that are conceptually related to selected document.

Reveal AI recommends in real time other documents to review when a document is selected.

Recommends documents based on over twenty different factors and will indicate why the documents are recommended.

Reveal AI recommends other entities to interrogate, including custodians, people, locations and topics.

Recommendation

Behavior Intelligence Modules

The Reveal AI behavior intelligence modules is based on finding text within emails that has words/phrases conveying a sense of pressure, opportunity, and/or rationalization (based on the fraud triangle theory).

Each behavior intelligence module is built based on a list-based classifier and a feature-based classifier. The latter includes:

  • Email metadata such as recipient count and subject line.

  • Linguistic characteristics such as personal pronouns ratio.

  • Statistical characteristics such as unique hits, ratio of sentences with hits and segment length.

Behavior Intelligence