> For the complete documentation index, see [llms.txt](https://cogneed.gitbook.io/cogneed-centre-daide/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://cogneed.gitbook.io/cogneed-centre-daide/cogneed-help-center/help/glossary.md).

# Glossary

### PWA (CogNeed Live) Features

* **Session**: When the user switches to Session mode, voice detections are activated for dynamic display of content cards on CogNeed AI™.
  * If CogNeed is integrated with client telephony, a session starts and ends automatically when a user makes a call.
  * If CogNeed is not integrated, users are advised to manually start and end a session for each call made.
* **Content card**: Typical CogNeed card gathering the content of a client on a given subject, which can be associated with 0 - N vocal commands.
* **Vocal command**: The trigger is a group of words that must be pronounced either by the agent or by the other person (Dual Channel). The trigger will only work if the words are pronounced as documented or very close (conjugation, 3 seconds instead of 2 seconds for example). The advantage of this mode is the speed of detection. It is your best option when you wish to induce your reps to use certain terms or messages.
* **Intention**: The trigger is a sentence spoken either by the agent or by the other person (Dual Channel) whose meaning is close to the documented intention, even if the terms used are different. The advantage is that it can respond to a wider range of expressions; the disadvantage is that the trigger is therefore less precise, and also a little longer because processing is done when the sentence is finished.
* **Detection**: Understanding by CogNeed AI™ that a vocal trigger was pronounced.
* **Display**: Occurs when a content card is displayed on a user screen : whether following a detection, a search or other actions.
* **Display (vocal triggered)**: Content card that will be displayed on the agent's screen when the associated vocal trigger is spoken.
* **Contextual display**: Adding a contextual display allows you to specify a particular display for your vocal trigger. It will only appear once the associated context, and then the vocal trigger, have been detected.
* **Text search**: Searching for information by writing in CogNeed AI™.
* **Decision Tree**: Ordered sequence of questions allowing to establish a diagnosis, to conduct a qualification, or to obtain a recommendation, based on a certain number of variables.
* **Link**: Link allowing to access complementary information in a new window (document, content sheet, decision tree, web page, etc).
* **Associated card**: Feature to connect two content cards together and facilitate the transition from one to the other.
* **Condensed content block:** Within a card, content initially presented by its title only, but which can be expanded in a single click by the user. These blocks, by being condensed, improve the readability of the content cards for users.
* **Call Compliance**: Compliance and communication rules that users must follow during their calls, subject to reminders in the Web Application. These rules are set up and documented in the CogNeed Dashboard and are subject to monitoring indicators.
* **Conversation Quality**: Rules regarding recommended and to be avoided formulations during calls. These rules are set up and documented in the CogNeed Dashboard and are subject to monitoring indicators.
* **Topics**: Conversation topics that may arise in calls, grouping one or more vocal triggers, associated with compliance and conversation quality rules and allowing for statistical analyses on formulations.
* **News**: Information element intended for all users of one or more projects and published on the homepage of the Web Application for a specified period.

### Dashboard Features

* **Dashboard**: The CogNeed web page allows for the configuration of the Web Application, management of content and vocal commands, and access to reports for authorised users.
* **Active/Inactive/Temporary Content card**: An active content card is available in the Web Application via voice or text search. A content card that is temporarily active becomes inactive when it reaches its expiration date. After this date, it remains in the knowledge base as an inactive content card, available to content managers but unavailable to users.
* **Content** **Library**: The content library contains content cards intended to be shared among several projects. The modification of these cards is done centrally and applies to all projects in which this card is published.
* **Publishing**: The action of making a library content card available in one or more projects, making it accessible to users of those projects.
* **Tags**: Element used to organise content cards. Each card can have one or several tags. Cards can be filtered by tag.
* **Reports**: Dashboard menu providing access to CogNeed reports on the Web Application usage.
* **Export**: Feature allowing export of CogNeed reports in the .csv format.
* **Advanced Analysis Item**: Feature allowing you to define a question to be automatically asked of the transcripts of upcoming CogNeed sessions. Advanced Analysis Items (AAIs) can be configured with execution conditions so that the question is only asked for sessions meeting certain criteria, for example:

  * duration greater than *X* minutes;
  * specific session status;
  * True/False answer to another AAI;
  * prior detection of a Context.

  Questions are asked via Generative AI models, applied only to transcripts with sufficient volume to reduce the risk of hallucination: sessions longer than one minute, containing at least 3 sentences and a minimum of 100 characters.

  For reporting and categorisation purposes, an AAI can be:

  * generic,
  * compliance-related,
  * quality-related.
* **Evaluation Sheet**: Feature allowing you to build a complete evaluation grid used to analyse CogNeed sessions. It can contain:

  * modules grouping several elements (AAI / compliance / quality),
  * standalone elements added outside of any module.

  The evaluation sheet allows you to:

  * assign a weight to each element;
  * define criticity levels;
  * once published in a project, automatically calculate users' scores across all their sessions.
* **Compliance Module**: Feature allowing you to categorise and group several elements:

  * AAIs,
  * real-time compliance elements,
  * real-time quality elements.

  The compliance module is used to structure the analysis around three main objectives:

  1. Build a readable evaluation sheet, by creating coherent sets;
  2. Group these elements into logical sections within reports;
  3. Organise the same elements in CogNeed Live, to make them easier to read in real time.
* **Phrasing Performance**: Feature allowing you to compare the effectiveness of several similar phrasings in order to identify the one that performs best with customers or prospects.

  It includes three components:

  1. Setting up the comparable phrasings (sentences to test);
  2. Intelligent exploration of transcripts to automatically detect similar phrasings spoken by users but not yet entered;
  3. Defining the "success" criterion, for example:
     * session duration after the phrasing is spoken,
     * final call status.

### Hierarchy

* **Organisation**: Entity incorporating all the information of a given client
* **Project**: Sub-division of an organisation that can be created for the needs of a project, campaign, team, etc.
* **CSM**: The Customer Success Manager (CSM) is a professional who works with our clients to ensure they are satisfied by helping them achieve their goals through the CogNeed AI™ solution.

### Tech

* **Progressive Web Application (PWA)**: A Progressive Web Application (PWA) is a web application designed to offer a user experience similar to that of a native app. It utilises modern web technologies to allow users to access a web application through a browser, while benefiting from advanced features such as offline operation, OS notifications, and installation on the device's taskbar.
* **LLM**: A Large Language Model (LLM) is an AI that understands and generates text by learning from a huge database of written language.
* **API**: An API (Application Programming Interface) is a set of rules that allows different software applications to communicate with each other.