This feature is now available as part of the Freddy Copilot add-on.

Assist bot, powered by Freddy AI, helps you create guides to onboard agents faster, ensures consistency in agent responses, boosts agent morale by enabling them to resolve tickets faster and with context. You can set up conditional flows inside Freshdesk for every issue type in your support.

The agents will be able to trigger the flows relevant to their agent-group through the Freshconnect widget and use it to provide consistent responses and faster resolution. Assist bot also understands which channel the agent is working on, like chat, call, or ticket to provide contextual assistance.


Bot Version

As you make changes or improvements to your bot, you can save them as versions. You can choose to publish a version so that it can start assisting agents and/or create a draft version to make additional changes for the next release while the current version is actively being used. Older versions can be archived so that change management is seamless.


Bot Conversation

A transcript of the actual conversation between an agent and a bot.



These are building blocks of a bot conversation. You can configure messages, set up conditions, get user input, and set up actions within dialogs.



A set of dialogs grouped together, create a flow. Within a flow, dialogs are sequentially connected to each other by default.


Conversation Break

The sequential processing of dialogs within a flow can be broken using the conversation break option. 


Default Flows

  • Hello - First flow when the agent returns to the bot. Has greetings dialog by default.

  • Sorry - Executed when the bot is unable to proceed further in the conversation. 


User Input

Configured for a dialog. This is the type of input required from the agent. Enabling the ‘Get user response’ toggle allows you to configure answer types. You can configure this manually or choose to define them dynamically.


Manual Answer Types

  • Button - For the agent to choose between options

  • Drop-down - For the agent to select an option from a list

  • Carousel - For the agent to select one card from the carousel

  • Text field - For the agent to type an answer in the text box

  • Date & time - For the agent to enter date and time from the calendar


Dynamic Answer Types

Answer types whose values are obtained from APIs. For example, this can be used when you configure buttons whose names are to be fetched through an API call. 


Text variants

Setting up variations of a message within a dialog can make your bot sound more human-like. A variation is displayed at random from the list you create.



The Assist bot’s builder allows you to set up dialogs and connect them to another, anywhere in the workflow. For every dialog you create, you can configure a follow-up action for the bot to perform. These actions can be any of the following:

  • Trigger API - This action will allow you to trigger APIs to fetch, update, or post data remotely

  • Add attribute - This action will enable you to add or update attributes of conversations or an agent profile

  • Make bot inactive - This action will make the bot inactive for a specific period of time

  • Start a new conversation - This action will enable you to end the current conversation and start a new one



A condition defines how the bot should progress in a conversation. It can be based on dialog inputs, API responses, custom parameters, placeholders, or function outputs. For example: After Dialog A is executed, proceed to Dialog B OR if Button C is clicked in Dialog A then proceed to Dialog D. 



The Assist bot’s builder provides you with commonly used functions for your daily operations, such as fetching current date, separating strings by a delimiter, manipulating date and time, etc. These Functions act as placeholders and take care of populating values for fields dynamically.

For example:

Split function:

This function splits the given text based on the delimiter/separator and returns the value based on the index specified.

Syntax: $fn{{#split(‘text’,’delimiter’,’position of output’)}}

Usage: $fn{{#split('chat&bots','&','1')}} 

The ‘&’ here is the separator which will split the text, 1 is the index indicating the position of which value should be returned (0 being the position of the first value).

Output: bots



The preview option allows you to test the bot based on its current configuration. 


API library

Configure and manage APIs used in the bot conversation. Once APIs are configured here, they can be triggered using the actions tab while configuring dialogs. 



This page will have a list of all the conversations between the Assist bot and the agents. You can apply the filters on this page to view just the list of conversations you need. 



When you have agents working from around the globe, a bot talking to them in a language they best understand will improve your agent efficiency and resolution rate. Within the Assist bot’s builder, multiple languages are configurable in a single pane, thus making your bot polyglot.


Quick actions

These are options shown to the agent to choose from, during the bot conversation. It comprises of slash commands, widget menus, and preset buttons. Preset buttons can be defined globally and are customizable for every conversation. Slash commands and widget menus can be defined globally but are not customizable for every conversation.


Intent-based flows

You can enable your Assist bot to understand the intent of your agents in a bot conversation. Set up questions and its variants and map them to an answer or a flow so the Assist bot can detect the intent and direct the agents to the right answer.


Group Mapping

Once you set up your Assist bots in the builder, map them to the required groups to provide the relevant assistance to your agents. For example: Map the Billing Assist bot to the Billing group in Freshdesk so that these flows appear to only the relevant agent-groups in Freshconnect. 

Trigger the Assist bot from Freshconnect

Your agents can start using the bots you’ve configured from the Freshconnect widget. Freddy understands which channel an agent is working on, like chat, call, or ticket, and allows the agent to quickly choose the appropriate bot for that channel and provides contextual assistance correspondingly.