Skip to main content

​
Overview

Zild Chat is a conversational application that enables users to interact directly with a generative AI model through a chat interface. It functions similarly to ChatGPT, but within the Zild Platform’s:
  • Tenant isolation model
  • Role-based access control
  • Integration framework
  • Audit and monitoring system
Zild Chat allows controlled AI usage within enterprise environments.

​
Architecture

Zild Chat operates through the following structure: User → Chat Interface → Agent → AI Model → Tool Execution (optional) → Response

​
1. User Interface

Users interact through a chat interface where they can:
  • Ask open-ended questions
  • Provide instructions
  • Upload content (if enabled)
  • Continue multi-turn conversations
The interface maintains session context.

​
2. Agent Layer

Each Zild Chat instance is connected to an Agent configuration. The Agent defines:
  • System instructions
  • Behavioral constraints
  • Tool access permissions
  • Knowledge sources
  • Escalation rules (if applicable)
This ensures controlled AI behavior.

​
3. AI Model Execution

Zild Chat connects to configured model providers such as:
  • Azure OpenAI
  • OpenAI
The model:
  • Processes user input
  • Maintains contextual memory
  • Generates natural language responses
  • Executes structured tool calls when required

​
4. Tool Execution (Optional)

Zild Chat can allow the AI model to execute tools such as:
  • Database queries
  • API calls
  • Document analysis
  • Structured workflows
Tool usage is strictly defined in the Agent configuration.

​
Core Capabilities

​
Multi-Turn Conversations

Zild Chat maintains conversation context across messages. This enables:
  • Complex reasoning
  • Iterative refinement
  • Follow-up questions
  • Contextual responses

​
Knowledge-Based Assistance

Zild Chat can be configured with:
  • Internal documentation
  • Policies
  • Product manuals
  • FAQ databases
This improves response accuracy and domain specificity.

​
Controlled AI Behavior

Unlike public AI interfaces, Zild Chat allows:
  • Restricted model capabilities
  • Defined instruction boundaries
  • Controlled tool access
  • Enterprise monitoring
This ensures predictable and compliant AI usage.

​
Enterprise Features

Zild Chat inherits platform-level capabilities:
  • Tenant isolation
  • Role-based permissions
  • Conversation logging
  • Audit trails
  • Integration control
  • Monitoring via Zild Insight
This makes it suitable for regulated or sensitive environments.

​
Typical Use Cases

Zild Chat can be used for:
  • Internal knowledge assistant
  • AI-powered documentation helper
  • Business analytics Q&A
  • Drafting content
  • Code assistance
  • Policy interpretation
  • Document review
It provides a secure AI workspace for enterprise users.