Graphlogic.ai platform

Registration and users

Registration

The first step is to create an account on the

@Platform

⁠

. To register or log into an existing account, use the

Registration and authorization

⁠

article.

Account and company settings

An

@Agent

⁠

is a

@Platform

⁠

object, the configuration of a future

@Bot

⁠

. A

@Bot

⁠

is an

@Agent

⁠

deployed to some

@End Channel

⁠

and communicating directly to

@Bot User

⁠

.

To familiarize yourself with the

@User Account

⁠

and

@Company

⁠

settings and find out what information is available for editing, see the

User account and Company

⁠

article. ​

⁠

⁠

Multi-user mode

The

@Platform

⁠

supports a multi-user mode, which enables users to invite each other to their

@Company

⁠

and collaborate on

@Bot

⁠

s. You will learn how to invite users to the

@Company

⁠

from the article

User account and Company

⁠

. ​

⁠

⁠

Users are granted particular permissions in accordance with their roles. To learn more about user roles, see the article

User Roles and User Management in Company

⁠

.

Subscription plans

The

@Platform

⁠

can be used on a paid or free basis. To familiarize yourself with the terms of subscription plans and learn how to switch between them, study the article

Subscription and Payment

⁠

.

Agents and Scripts

Platform tabs (top bar)

Now let's take a look at the

@Platform

⁠

tabs. They are available on the top panel: ​

⁠

⁠

To learn more about the tabs, see

Overview of Platform Tabs

⁠

.

Creating Agents

⁠

@Agent

⁠

s are created on the

@Project

⁠

s tab. For detailed information about the tab, see:

Dashboard (projects) tab

⁠

.

Information on creating an

@Agent

⁠

:

Creating, editing and deleting an Agent

⁠

.

Information on exporting, importing, cloning and replacement of the

@Script

⁠

:

Export, import, cloning and replacing an Agent

⁠

.

Available step-by-step instructions on

@Agent

⁠

creation: ​

Creating a Simple Survey Bot

⁠

⁠

Creating a Bot with Speech Recognition

⁠

⁠

Creating a Bot with Weather Service Integration

⁠

⁠

Botbuilder

The

@Agent

⁠

@Script

⁠

is created in the

@BotBuilder

⁠

. You can find out how to switch to

@BotBuilder

⁠

and learn about its functionality in the

Overview of BotBuilder and its tabs

⁠

article. ​

⁠

⁠

Slot types

After familiarizing yourself with

@BotBuilder

⁠

and its tabs, it is essential to study

@Slot

⁠

s which comprise the

@Agent

⁠

@Script

⁠

. Each

@Slot

⁠

corresponds to a specific action of the

@Agent

⁠

. For detailed information on the types of

@Slot

⁠

s, as well as instructions on how to create and utilize them within the

@Script

⁠

, refer to

Creating and types of Slots and Subslots

⁠

.

For information on how to move and clone

@Slot

⁠

s and

@Script Branch

⁠

es see

Moving and cloning Slots and Script Branches

⁠

.

For information on how to make it easier to navigate the

@Script Tree

⁠

by collapsing the

@Script Branch

⁠

es and coloring the

@Slot

⁠

s:

Script tree navigation

⁠

.

Articles elaborating on settings and operating features of each

@Slot

⁠

are available in the

How to create and configure Slots

⁠

section.

Natural language recognition by Agents

⁠

@Agent

⁠

s can interpret natural language contained in

@Bot User

⁠

‘s messages.

@Agent

⁠

s can recognize both the entire intention of the

@Request

⁠

(

Natural language recognition

⁠

) and a separate entity (

Recognizing entities | Vocabularies

⁠

).

Chat context variables

For further insights into the

@Context Variable

⁠

s, see the

How to use Chat Context in your Agent

⁠

section.

Agent training

To communicate with the

@Agent

⁠

, after its

@Script

⁠

is fully assembled, it needs to be trained. For successful

@Agent

⁠

@Training

⁠

, it must meet the requirements described in the article

Requirements for successful Agent training

⁠

. This article will also guide you on how to resolve any errors that may occur during the training process.

Agent testing

To learn how to properly test the

@Script

⁠

of the

@Agent

⁠

and improve its natural language understanding, you can refer to the articles

Testing the Agent script

⁠

and

Testing and improving Agent recognition: selection of Confidence threshold

⁠

.

To improve natural language understanding in the

@Agent

⁠

, you can add new

@Training Phrase

⁠

s to its

@Training Dataset

⁠

directly on the Analytics tab:

Additional training of the Agent from the Analytics tab

⁠

.

Channels and integrations

Deployment of Agents to Channels

After the

@Agent

⁠

@Script

⁠

has been tested, the

@Agent

⁠

can be deployed to

@End Channel

⁠

s for communication with

@Bot User

⁠

s. ​

⁠

⁠

For detailed instructions on deploying

@Agent

⁠

s to

@End Channel

⁠

s see section

How to deploy your Agent to channels

⁠

.

Chat API

The

@Agent

⁠

can also be deployed to

@End Channel

⁠

s that do not have a built-in

@Connector

⁠

in the

@Platform

⁠

. For this purpose, there is a universal

@Connector

⁠

called Chat API with a public API. For more information, please refer to

Chat API

⁠

.

Custom integrations with external systems

The

@Platform

⁠

implements a feature that allows to integrate the

@Agent

⁠

with external systems that provide an open API, such as CRM. This can be done using an

@External Request

⁠

. For more information, please refer to

External Request slot (Sending Requests to APIs from script)

⁠

.

Additionally, there is a feature that enables sending a request to the

@Agent

⁠

from an external system, allowing the

@Agent

⁠

to initiate a conversation with the

@Bot User

⁠

. This functionality can be used, for instance, to send notifications to

@Bot User

⁠

about the status of their order. This can be achieved using the

@Incoming Request

⁠

@Slot

⁠

. For more information, please refer to

Mailings and notifications | Incoming Request slot

⁠

.

WhatsApp notifications

The

@Platform

⁠

implements a feature that allows sending notifications from the

@Agent

⁠

to the

@Bot User

⁠

through WhatsApp. This is done using the

@Notification

⁠

@Slot

⁠

, which is designed for sending messages when a

@Chat

⁠

with the

@Bot User

⁠

has not been created yet, or when the

@Agent

⁠

is unable to initiate a conversation after a specific timeout (due to limitations of certain channels). For more information, please refer to

Mailings in WhatsApp | Notification slot

⁠

.

Bot work and communication

Limits

There are limits set for certain events in the

@Platform

⁠

. You can learn more about these limits in the following article:

Limits

⁠

.

Communication analytics

Within the

@Platform

⁠

, you can access analytics through the

Analytics tab and filters

⁠

, as well as download statistics detailing communication between

@Agent

⁠

s and

@Bot User

⁠

s. ​

⁠

⁠

History of communication between

@Agent

⁠

s and

@Bot User

⁠

s:

Chat history

⁠

⁠

⁠

@Context Variable

⁠

s of all

@Chat

⁠

s:

Chat context report

⁠

The number of times a particular

@Slot

⁠

has been executed by an

@Agent

⁠

:

Slot stats report

⁠

Information on the

@Slot

⁠

s which were last executed within

@Bot User

⁠

dialogs with a particular

@Agent

⁠

(i.e. the

@Slot

⁠

s which mark the end of dialogs):

Stop slots report

⁠

Information about how many times a certain

@Slot

⁠

was executed for the selected period of time and general statistics on the

@Agent

⁠

:

Agent usage

⁠