AI Features in Bigin

AI Features in Bigin

Info
Early Access
AI features in Bigin are currently in early access and will be rolled out to all users over the next few weeks.

The AI features in Bigin are designed to help you save time, improve communication clarity, and stay focused on your core business activities.

These AI features are categorized into two types in Bigin:
  1. Zia Assistive AI: Features that require user intervention
  2. Zia Agents: Fully automated digital employees that work in the background
By bringing Zia Assistive AI and context-aware support directly into Bigin, you no longer need to switch between external tools or platforms to draft emails, summarize information, or gather insights. Everything you need is available within your organization. On the other hand, Zia Agents function as AI-powered digital employees that are designed to analyze interactions and automatically perform actions based on predefined rules. For example, they respond to emails, evaluate deal activities, or trigger follow-up actions automatically. Built with Zoho Agent Studio, these agents operate independently in the background to reduce workload and improve productivity.

For example, sales teams often spend time reviewing emails, responding to inquiries, analyzing deal losses, and identifying upsell opportunities. Zia Agents streamline these tasks by performing intelligent analysis and automated actions without user intervention. This enables sales teams to focus on closing deals while the AI handles repetitive work.

Availability

Info
All AI features are available in the Bigin Premier and Bigin 360 editions.

Credits

Each plan includes a monthly allocation of AI credits.

Edition
Monthly AI Credits
Premier Plan
1000
Bigin 360
3000

AI credit consumption varies based on the feature used and complexity of the request.

You can view your AI credit usage by navigating to Settings > AI.


Zia Assistive AI - Instant help in Bigin

Zia Assistive AI features work automatically and help improve efficiency without requiring human intervention. Bigin's Zia Assistive AI features are listed below.

Writing Assistant for Email Composer

The email composer includes Zia’s writing assistant to help you draft, refine, or improve email content. You can click the Ask Zia button in the email composer to open a chat window, describe the type of email you need, and let Zia generate a complete message, including the email's subject.



You can customize the style and content length by clicking the settings icon and selecting your preferred options. Once configured, any prompt you provide will generate content that follows your chosen preferences.


To improve the existing content, selecting text inside the composer displays a Zia icon, which offers options to rephrase, adjust tone, translate content, shorten text, or apply other improvements.


Proofread Email

Zia Assistive AI can proofread your content and provide recommendations to improve it. A dashboard is displayed showing suggestions like spelling mistakes and grammar recommendations, along with a readability score for better understanding of the email content.


Record Summary

Zia Assistive AI can summarize your contact, company, and pipeline record details on demand. This helps you quickly understand the current status of  a record without reviewing every detail manually.

How it works
  1. Open any contact, company, or pipeline record.
  2. Click Summarize.
  3. Zia analyzes the entire record, including activities such as emails, notes, and tasks.
  4. A concise summary is generated in the background.
Once the summary is ready, the Summarize button changes to Show Summary. You can click this button to view or copy the summary.
The summary will provide you with an overview and key insights about the record, along with recommended actions.


Notes Summary

Zia Assistive AI can summarize the most recent notes added to a record. It processes the latest ten notes and presents a clear, easy-to-read summary, helping users understand the context quickly.


Email Summary

Inside the Email tab, Zia Assistive AI summarizes entire email conversations, including Email-In threads. This allows users to quickly review lengthy email chains without manually reading through each message.


WhatsApp Summary

Zia Assistive AI analyzes the last ten WhatsApp messages exchanged with a customer and generates a summary that captures the key points of the conversation. This is especially useful for long chats or scattered messages.


Zia Agents: AI-powered digital employees

In addition to Zia Assistive AI features, we also have Zia Agents. Zia Agents are designed to analyze interactions and automatically take action based on predefined rules.

Configuring Zia Agents

All Zia Agents follow a unified setup flow. To get started:
  1. Navigate to Settings > AI > Agents.
  2. Select the agent you want to enable and click Try this Agent.
  3. You will be redirected to the authorization screen, where you need to authorize Bigin to allow the agent to interact with your data.
To set up a Zia Agent, you need to create a Digital Employee with the appropriate permissions. During configuration, you must assign a role and profile for the agent.

Once the digital employee is created, Bigin automatically generates a system email address for it. All actions performed by the agent will appear under this email in the audit logs.



Notes
Notes:
It is recommended to select Level 1 and Administrator to allow the agent required access to analyze records and perform automated tasks.


After completing the authorization, you will be asked to create a user for the AI Agent. All actions performed by the AI Agent will be executed through this user, which acts as a digital employee in your Bigin account.

To set this up:
  1. Provide a name for the digital employee
  2. Assign the required roles and profiles. It is recommended to select Level 1 and Administrator to allow the agent sufficient access to analyze records and perform automated tasks.
  3. Once the digital employee is created, Bigin automatically generates a system email address for it. 
  4. You can manage the AI Agent user’s profile and permissions from Users > AI Agents when needed.


This section lists the digital employees created for each agent. If a digital employee is deactivated, all associated Zia Agent functionalities will stop.

Notes
Notes:
  1. If the digital employee created for an agent is deactivated, the agent will immediately stop functioning.
  2. All actions performed by the agent are logged under the digital employee’s system-generated email address inside the audit logs.

Zia Agents

Once the digital employee is created, you will be redirected back to Bigin to configure the agent rules.
Bigin offers multiple Zia Agents, each designed for a specific business purpose. While all agents follow the same installation process, their rules and actions differ based on the outcome they are built to achieve.

Bigin's Zia Agents are listed below:
  1. Churn Analyzer
  2. Cross-Sell Genie
  3. Reply Assistant for Email-In 
These agents continuously monitor records or communication channels and perform automated actions based on configured rules.

Churn Analyzer  

The Churn Analyzer automatically triggers whenever a deal is moved to the Closed Lost stage. When this happens, the agent reviews past interactions, including emails, notes, calls, and activities, to identify the most probable reason behind the lost deal. Once the analysis is complete, the agent sends these insights directly to the deal owner, helping sales teams understand customer concerns and refine their sales approach.

To create a rule under Churn Analyzer, click + Create Rule and select the required pipeline and sub-pipeline where the rule should apply. Based on the option you choose, the identified reason for loss will either be added as a note on the deal or sent as an email to the respective record owner.



For example, imagine a deal for Zylker Technologies managed by Amelia Burrows. After several follow-ups, the customer stops responding, and Amelia marks the deal as Closed Lost. The Churn Analyzer then reviews the deal’s emails, notes, and activities and finds that the customer had earlier mentioned pricing concerns.

Instead of Amelia going through all the interactions themselves, the Churn Analyzer identifies pricing issues as the likely reason for losing the deal and sends this insight directly to Amelia. This helps them quickly understand what went wrong and improve their approach for future deals.

Once a rule is executed, the resulting insights will appear directly inside the deal. Below is an example of how the Churn Analyzer displays the identified reason for loss.



You can view and manage all your rules by navigating to Setup > AI > Churn Analyzer > Manage. From here, you can:
  1. Edit an existing rule by clicking the edit icon.


  2. Disable a rule by turning off the status toggle.


  3. Delete a rule if it is no longer required.


Cross-Sell Genie

The Cross-Sell Genie identifies potential cross-selling opportunities whenever a deal is marked as Closed Won. The agent reviews the product associated with the deal and predicts additional items the customer may be interested in based on product relevance and purchasing patterns.

To create a rule under Cross-Sell Genie, click + Create Rule and select the pipeline and sub-pipeline where the rule should apply. The agent’s recommendation can be emailed directly to the customer, or you may choose to save the email as a draft so you can review and send it manually.



You can further improve the accuracy of these recommendations by uploading a document as a Knowledge Source. To do so, click + Knowledge Source and the Upload Files button, or simply drag the file into this window. This document can include detailed information about your products and the relationship between them. The agent will reference this content to generate more precise cross-sell or up-sell suggestions.

You can add selected fields from the Product module as a Knowledge Source. During configuration, data from the Active Products custom view is added to the knowledge base. After setup, any new products or updates to existing products are automatically synced every hour, ensuring the information stays up to date. The agent uses this knowledge base to recommend relevant products to customers.



For example, if a customer purchases an air conditioner, the Cross-Sell Genie may recommend related items such as a voltage stabilizer, AMC service package, or air purifier. These suggestions are based on the product details and relationships defined in your uploaded Knowledge Source. The agent analyzes the primary product purchased, identifies complementary items, and generates an email containing these recommendations. Depending on your configuration, the email can either be sent automatically to the customer or saved as a draft for your review.

Once the rule is executed, the recommended products will be displayed within the deal record. The screenshot below illustrates how the Cross-Sell Genie presents these suggestions.



You can view and manage all your rules and knowledge source that you create. To do so, go to Setup > AI > Cross-Sell Genie > Manage. From here, you can:
  1. Edit an existing rule by clicking the edit icon.


  2. Disable a rule by turning off the status toggle.


  3. Delete a rule if it is no longer required.


Reply Assistant for Email-In

Prerequisites

If Email-In is not integrated, this agent will not work. You will be asked to integrate Email-In to activate the agent. You can click the Enable Now link to be redirected to the Email-In installation page.



The Email-In Reply Assistant helps your team manage incoming emails more efficiently. Once Email-In is set up and the agent is enabled, it automatically reads every new email, understands the context, and generates an appropriate reply.

To create a rule under the Email-In Reply Assistant, click + Create Rule and select the Email-In address the rule should apply to. You can configure the agent to add the reply as a comment, send a direct response to the customer, perform a spam check, or automatically tag the email for easier categorization.



You can also control when the agent should reply. The assistant can respond immediately, or you can define a time interval before the reply is sent. Additionally, you may choose not to send automated replies when the user is online, ensuring the agent responds only when the user is offline. A disclaimer can also be added to inform customers that the message is auto-generated.



You can further improve the accuracy of these reply emails by uploading a document as a Knowledge Source. The agent will reference this content to generate more precise reply emails.



For example, if the support address support@zylker.com receives an email from a customer requesting a copy of their invoice, the Email-In Reply Assistant checks the Knowledge Base you have uploaded, for instance, a document describing billing processes, invoice retrieval steps, or payment details. If the assigned salesperson is offline, the assistant analyses the customer’s message and drafts a response using the information from the Knowledge Base.

If the reply delay option is configured, the agent will wait for the specified interval before sending the message. Once the rule executes, the response is either sent automatically or saved as a draft for the user to review, depending on the chosen settings.



If the received email is spam, the agent will detect it and add a comment stating that the email is spam.



Similarly, if the email contains details about any activities to be scheduled, those details will also be added as a comment.



You can view and manage all your rules and knowledge source that you create. To do so, go to Setup > AI > Email-In Assistant > Manage. From here, you can:
  1. Edit an existing rule by clicking the edit icon.


  2. Disable a rule by turning off the status toggle.


  3. Delete a rule if it is no longer required.



Notes
Notes:
  1. Emails received before the rule was created will not be processed by the agent. The Email-In Reply Assistant will only respond to new conversations that start after the rule is set up.
  2. If the agent is unable to generate an appropriate response, it will automatically @mention the record owner in a note, prompting them to review and handle the conversation manually.