Concepts and Components

The section describes key concepts of the Session AI Platform, which are important to understand before using the platform.

Namespace

The Session AI Platform has a dedicated account for each business, which is called a Namespace, where the business creates its use cases and receives all incoming customer data. The events go into their namespace as per the API key you use for the namespace. This namespace is given a type according to the business category/domain. For example, for a retail business, its namespace category/domain will be Retail, and all configuration for a particular business account is created within its designated namespace.

Session AI Platform

The Session AI Platform is a web-based application. It allows you to create and configure the required metadata and configuration to build and execute a business's use cases.

Source Channels

Source Channels are also called Data channels or Event Sources, which are sources of incoming customer data into the Session AI Platform. Session AI SDKs gather customer data from various sources or channels and send them as events to the Session AI platform. The platform uses API keys to authenticate all incoming traffic. Examples of source channels are mobile phones, desktops, Webhook receivers, in-store beacons, etc.

Experiences

The main focus of Session AI is to enable businesses to present meaningful and relevant information to the right customer at the right time, which leads to increased customer engagement and conversions across all channels.
Session AI provides this capability using experiences that allow businesses to send actions to users without requiring them to write code. The basic components of the experience are:

• To Whom the actions are sent to (user segment.)
• When the actions are sent (based on events or schedule.)
• What actions are sent (push notification, in-app message, etc.)

Template

A template is a group of Session AI components packaged together that is used to create an experience. Business Users can customize a template and execute it to deliver experiences to the users.

Session AI offers four types of templates - Experience, Segment, Trigger, and Action templates. A Business User selects a template, customizes it as per the business requirements, and executes it to deliver experiences to customers.

Marketplace

The marketplace is a repository that contains pre-built templates for various domains such as Retail, Hospitality, Travel, etc. Business Users can subscribe to these templates and use them as per their requirements.
Templates are published either to the marketplace or to a namespace.

• Marketplace Templates: These templates are published to the marketplace and are generic in nature. They can be subscribed and configured by Business Users as per the requirement of the use case.
• Namespace Templates: These templates are published to a namespace.

Events

An event is a meaningful action or an activity, which the customer performs on a channel using any application. For example, login activity, add to cart, pay the bill, etc.
Note: Events need to be first created in the Session AI Platform before they can be recognized in the incoming event.

The incoming event has an event name and some default parameters in key-value format (Parameter Name: Parameter Value). The default event parameters are sent with the event, and they provide the necessary context about the event.

For example, the default parameters for an event called Login:

• Source: SDK
• Source Channel: Phone
• Source OS: Android

Domain Events

Events belonging to a particular domain like retail and are defined and registered by Session AI when the Namespace account is created. They represent typical activities that customers do in the business source channel and are used by Session AI for creating any out-of-the-box configuration for that domain. Domain event names always start with the “z1_” prefix . For example, domain event for login would be ‘z1_login”.

Namespace Events

These are the events coming into the Session AI platform from any source channel. These events should be defined within the Session AI platform before the platform can recognize them as valid events coming into a designated namespace.

For each incoming event, the Session AI platform provides the Event Enrichment section that allows you to enrich them using code snippets and methods.

Event Mapping

The template published from the marketplace can be used only if they are mapped with your namespace. Event mapping is a mandatory step for mapping your experience template's domain event to the right namespace event(s) which will trigger the instance of that experience template.

Customer Profile

A customer profile is a collection of information about an individual customer. The platform extracts customer information from incoming events and creates unique customer/user profiles for the customers.

Note: Custom user attributes need to be first created in the Session AI platform before they can be recognized in the incoming event.

The incoming events also have custom user attributes in key-value format (Attribute Name: Attribute Value), which provides additional context on the user who performed the event and therefore must be thoughtfully collected by the SDK.

For example, an event named “login” can have the following user attributes:

• Name: Jack
• Membership: Gold
• City: San Jose
• Device ID: 12344565789
• Location: Fremont, CA

Segments

In Session AI, a segment is a group of users who match the predefined criteria. The criteria can be based on the users':

• In-app behavior (for example, a user's activity pattern)
• Device attributes
• Customer attributes
• Interactions performed by the user

Segments are dynamic. As events from users come into the platform, their attributes are evaluated, and the customers who fulfill any segment criteria are added or removed from the appropriate segment dynamically.

Once you have defined a segment, the segment can be used in the experience. The users of the selected segment are targeted with the defined experience.

Triggers

Trigger is a configuration in the Session AI platform that is used to initiate real-time actions, and it is defined by a customer activity or other trends internal or external to the system. Triggers can also include events and user attributes. Session AI executes the trigger whenever a user performs the triggering activity, and the executed trigger kicks off the execution of the attached downstream workflow, i.e., an associated experience.

Actions

Actions are messages or information you want to send to the customers.
There are two categories of actions:

• Device notifications: Device notifications push notifications.
• App notifications: App Notifications include in-app messages such as full-screen messages, banners, alerts, AppBox messages, custom actions, etc.

Session AI provides you with out-of-the-box action templates, which allow you to quickly create actions using pre-configured content and layout.

A/B Testing

A/B Testing compares two action variants of the same experience to see which action best increases the conversions.

Optimized Allocation A/B Testing

An Optimized Allocation A/B Testing identifies a winner variant among two or more actions of the experience. This automatically shifts more traffic to the better performing action.
Advantages

• Identify a statistically significant winner faster than the A/B test
• Gives better conversions than A/B test

Success Metrics

Success metrics include key business metrics which define the success of a given experience. In Session AI, the success metric is configured within a time window that is set through the system configuration at the namespace level.
For example:
In Retail/E-commerce, Conversion represents purchases and it depends on the occurrence of a “placed order” event. Let's assume we have set Success Metric Attribution Window to 15 mins. If we send an experience with an offer and the "place order" event comes within these 15 mins, it is counted as a "Success".

Following are the OOTB success metrics available only for the "retail" domain

• Conversion
• Conversion rate
• Average order value
• Revenue per visitor
• Total revenue