Articles in this section
Generative AI Designer
Feature Availability
Vtiger Editions: One Growth | One Professional | One Enterprise | One AI
Introduction
Artificial Intelligence (AI) enables systems to simulate human intelligence such as learning, reasoning, problem-solving, and decision-making. Generative AI is a subset of AI that can create original content such as text, summaries, recommendations, or responses based on user prompts.
A Generative AI Designer in Vtiger CRM provides a centralized platform to configure, monitor, and improve AI-powered interactions across the CRM. It helps organizations manage bot behavior, train AI models using CRM data, analyze performance, and ensure secure handling of sensitive information. By bringing setup, analytics, and query management into a single module, Generative AI Designer enables businesses to deliver accurate, consistent, and intelligent responses to users and customers.
Benefits of Generative AI Designer
The benefits of Generative AI Designer are:
- Improved Customer Experience: Bots can respond quickly and accurately to customer queries, reducing response time and improving satisfaction.
- Centralized AI Management: All AI-related configurations, analytics, and queries are managed from a single module.
- Consistent and Reliable Responses: Training AI models using FAQs, articles, and documents ensures consistent answers across channels.
- Actionable Insights: Built-in analytics help track bot performance, query trends, and user interactions.
- Enhanced Productivity: AI agents and prompts reduce manual effort by automating routine tasks and responses.
- Data Privacy and Security: Sensitive information is protected using built-in data masking mechanisms.
Key Terminology
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Generative AI Designer Features
The Generative AI Designer is divided into four main sections, each serving a specific purpose:
- GenAI Setup – Configure how Calculus AI works
- GenAI Analytics – Analyze bot performance using visual reports
- GenAI Queries – Review AI queries and responses
- GenAI Token Usage – Track AI usage and consumption
These sections work together to help you build, monitor, and optimize AI-powered experiences in Vtiger CRM.
GenAI Setup
The GenAI Setup section is where all AI components are configured. This is the first and most important step when setting up Generative AI in Vtiger. Everything configured here determines:
- What data AI can access
- How AI responds
- Where AI is used
GenAI Setup includes the following interconnected components:
- RAG Models
- Agents
- Prompts
- Chatflows
- Data Shield
Let us understand each component in detail.
RAG Models
RAG (Retrieval-Augmented Generation) Models enable AI to retrieve relevant CRM records and documents before generating a response. Instead of answering purely from its language model, AI first retrieves trusted information from your CRM.
Why RAG Models are important:
- Ensures answers are based on real CRM data
- Improves accuracy and relevance
- Ideal for knowledge-based responses
How RAG Models work:
- A user asks a question.
- The RAG model searches selected CRM modules or documents.
- Relevant data is retrieved.
- AI generates a response using the retrieved information.
Examples:
- Cases RAG Model: Finds similar cases while handling support tickets
- Articles RAG Model: Generates answers from knowledge base articles
- FAQs RAG Model: Responds to frequently asked questions
Admins can define which modules or documents are included in the retrieval scope and monitor usage and performance.
Agents
AI Agents are intelligent assistants designed to perform specific tasks inside the CRM. Each agent is trained on defined topics and allowed to perform specific actions.
Why Agents are useful:
- Automate repetitive tasks
- Guide users step by step
- Perform actions directly inside CRM
How Agents work
An agent consists of:
- Purpose – What the agent is designed to do
- Topics – Areas the agent is trained on
- Actions – Tasks the agent can perform
Example: Case Agent
Topics:
- Case creation
- Case summary
- Case replies
Actions:
- Ask users for required details
- Create cases using CRM APIs
- Summarize existing cases
Calculus AI can automatically suggest suitable topics and actions based on the agent’s purpose.
Types of Agents:
- Customer-facing agents – Exposed through chatflows
- Internal agents – Operate within the CRM for employee assistance
Prompts
Prompts are instructions that guide how AI generates responses. They define the tone, structure, and context of AI-generated content.
Why Prompts matter:
- Control response quality
- Ensure consistent tone
- Improve predictability
What you can do with Prompts:
- Create prompts for summaries, emails, and reports
- Test and refine prompt outputs
- Reuse prompts across agents and chatflows
Examples:
A prompt can instruct AI to:
- Generate a polite customer support response
- Summarize a long case description
- Draft a professional email update
Well-crafted prompts lead to reliable and accurate AI responses.
To learn more about the Prompts, click here.
Chatflows
Chatflows allow you to design guided conversational experiences using a visual, no-code interface. These are mainly used for customer-facing interactions.
Why Chatflows are useful:
- Guide users through multi-step processes
- Collect structured data
- Improve chatbot experience
How Chatflows work
Each step in a chatflow can:
- Ask a question
- Trigger an AI response
- Use prompts or agents
- Interact with CRM data
Example:
A chatflow can guide a website visitor to:
- Describe an issue
- Capture required details
- Automatically create a support case
Chatflows can be deployed on websites and messaging apps such as WhatsApp. Only one chatflow can be active per channel.
To learn more about the Chatflow, click here.
Data Shield
Data Shield ensures sensitive data is protected before being sent to AI models.
Why Data Shield is important
- Maintains compliance
- Protects customer privacy
- Prevents exposure of sensitive data
How Data Shield works
It masks data that matches specific patterns such as:
- Email addresses
- Phone numbers
- Social Security Numbers
Masking happens before the data is processed by AI.
Understanding the GenAI Setup Workflow
All GenAI Setup components are interconnected.
Example Flow
- Create a RAG Model.
- Choose Articles or FAQs as the data source.
- Add the RAG Model to a Prompt.
- Define how AI should use retrieved data.
- Use the Prompt in a Chatflow.
- Add an Agent (optional).
- Customer-facing (mapped in chatflow)
- Internal (works within CRM)
This flow ensures AI responses are accurate, guided, and context-aware.
Training and Testing AI
After configuring GenAI Setup, AI must be trained and tested.
- Training occurs when FAQs, articles, or prompts are updated.
- Use Test Query to validate AI responses.
- Review analytics and queries to identify improvement areas.
- Retrain AI based on feedback.
This continuous loop improves response quality over time.
GenAI Analytics
The GenAI Analytics section displays visual reports and graphs showing how your AI bot performs over time.
You can:
- Track the number of bot queries
- Analyze trends and patterns
- Identify areas for improvement
These insights help refine prompts, retrain models, and improve AI accuracy.
GenAI Queries
The GenAI Queries section displays all AI-handled queries and responses.
You can:
- View accepted, rejected, or imperfect responses
- Identify queries needing improvement
- Add or update FAQs
This feedback loop helps continuously enhance AI performance.
GenAI Token Usage
The GenAI Token Usage section tracks AI consumption.
You can:
- See token usage per query
- Identify who initiated the query
- Monitor usage trends
This helps manage AI costs and optimize usage efficiently.
Accessing the Generative AI Designer
This section explains the prerequisites, location, and role-based permissions required to access and manage the Generative AI Designer in Vtiger CRM.
Prerequisites
Before accessing the Generative AI Designer, ensure the following:
- The Calculus AI Add-on is installed from the Vtiger Extension Store.
- You are logged in with an Admin profile to configure and train AI components.
Note: Non-admin users can interact with AI features but cannot configure or train AI models
Follow these steps to locate the Generative AI Designer in Vtiger CRM:
- Log in to Vtiger CRM.
- Go to the Platform module.
- Click Designers > Generative AI Designer.
The Generative AI Designer module opens, displaying the GenAI Setup, GenAI Analytics, GenAI Queries, and GenAI Token Usage sections.
Roles and Permissions
Access to Generative AI Designer features depends on user roles:
- Admin Users
- Configure GenAI Setup components such as RAG Models, Agents, Prompts, Chatflows, and Data Shield
- Train and retrain AI models
- Review AI queries and analytics
- Monitor AI token usage
- Non-Admin Users
- Interact with AI features based on assigned role permissions
- Submit AI queries through supported channels
- View AI-generated responses
- Cannot modify AI configurations, training data, or system-level settings
Reference
Check out the following links for related information: