Skip to main content

Introduction

Gentic is an advanced Agent Builder platform that enables you to create intelligent, knowledge-powered agents with customizable workflows. Our platform combines three powerful components to help you build sophisticated AI solutions:

Core Components​

1. Intelligent Agents​

Agents in Gentic are autonomous AI entities that can:

  • Process and understand natural language with advanced LLM capabilities
  • Make context-aware decisions and take appropriate actions
  • Maintain conversation state and context across interactions
  • Learn and adapt from user interactions and feedback
  • Execute complex tasks through workflow orchestration

2. Knowledge Graphs​

Our knowledge graph technology enables agents to:

  • Connect and integrate multiple data sources (databases, APIs, documents)
  • Build semantic relationships between different pieces of information
  • Perform intelligent search and retrieval across connected knowledge
  • Maintain context and relevance in conversations
  • Support real-time knowledge updates and synchronization

3. Workflow Engine​

The workflow engine allows you to:

  • Define complex, multi-step processes with conditional logic
  • Orchestrate parallel and sequential task execution
  • Implement error handling and retry mechanisms
  • Monitor and debug workflow execution
  • Create reusable workflow templates

Building Your First Agent​

  1. Create an Agent

    const agent = await client.createAgent({
      name: 'Customer Support Agent',
      description: 'Handles customer inquiries and support tickets',
      capabilities: ['conversation', 'ticket_management'],
      model: 'gpt-4', // Choose your preferred LLM
      temperature: 0.7 // Control response creativity
    });

  2. Build Knowledge Graph

    const knowledgeGraph = await agent.createKnowledgeGraph({
      name: 'Support Knowledge Base',
      sources: [
        {
          type: 'database',
          config: {
            connectionString: 'postgresql://user:pass@localhost:5432/support'
          }
        },
        {
          type: 'document',
          config: {
            path: './docs/support-guides'
          }
        }
      ],
      relationships: {
        'ticket_type': ['related_to', 'solution'],
        'product': ['has_issue', 'requires_fix']
      }
    });

  3. Define Workflows

    const workflow = await agent.createWorkflow({
      name: 'Ticket Resolution',
      steps: [
        {
          type: 'classify',
          input: '{{user_message}}',
          categories: ['billing', 'technical', 'account'],
          confidence_threshold: 0.8
        },
        {
          type: 'search',
          query: '{{classification}} {{user_message}}',
          source: 'knowledge_graph',
          max_results: 5
        },
        {
          type: 'generate',
          prompt: 'Based on the search results and ticket type, provide a solution',
          context: {
            previous_messages: '{{conversation_history}}',
            user_profile: '{{user_profile}}'
          }
        },
        {
          type: 'action',
          name: 'update_ticket',
          params: {
            status: 'resolved',
            solution: '{{generated_solution}}'
          }
        }
      ],
      error_handling: {
        retry_count: 3,
        fallback_action: 'escalate_to_human'
      }
    });

  4. Test and Deploy

    const response = await agent.converse({
      message: 'I need help with my billing',
      context: {
        user_id: '123',
        previous_messages: [],
        user_profile: {
          plan: 'premium',
          subscription_date: '2023-01-01'
        }
      },
      workflow: 'Ticket Resolution'
    });

Key Benefits​

  • Rapid Development: Build sophisticated agents in minutes, not months
  • Scalable Knowledge: Connect and manage multiple data sources seamlessly
  • Flexible Workflows: Create complex processes with visual workflow builder
  • Enterprise Ready: Security, monitoring, and deployment tools built-in
  • Continuous Learning: Agents improve over time with feedback and new data

Next Steps​