Building an Autonomous AI Travel Planner: Multi-Agent Orchestration & Real-Time API Tool I

About this event

Building an Autonomous AI Travel Planner with Multi-Agent Orchestration

Designing Personalised, Real-Time Travel Solutions with Agno, Gemini, and SerpAPI

Planning travel today remains a fragmented and time-consuming manual effort. Travelers routinely jump across flight comparison sites, hotel booking portals, and travel blogs—all while struggling to enforce specific dietary preferences (such as Vegan, Jain, or Halal), preferred cuisines, budget limits, or aesthetic, Instagram-worthy destinations. Generic LLM prompts often fail to synthesize live pricing data with complex personal constraints.

Moving beyond single-prompt chatbots, this session explores how to design an end-to-end Multi-Agent AI Travel Planner. By delegating tasks across specialized autonomous agents using the Agno Framework, integrating SerpApiTools for real-time web retrieval, and using Gemini as the reasoning core, we show how to deliver fully personalized, structured $N$-day itineraries alongside live booking resources.

In this session, you'll learn:

• Solving Travel Friction: Overcoming information fragmentation, manual synthesis overhead, and generic AI hallucinations with agentic architectures.
• Multi-Agent Orchestration: Implementing the Orchestrator-Workers (Planner-Worker) pattern to split complex travel workflows into specialized worker tasks.
• Real-Time Data Grounding: Integrating SerpApiTools (Google Flights Engine & Google Search API) to feed live ticket pricing and booking tokens directly into the LLM context.
• Hyper-Personalized Curation: Prompting worker agents to strictly validate granular dietary restrictions (Jain, Halal, Vegan, Vegetarian), preferred cuisines, and aesthetic/Instagram-popular photo spots.
• Chained Reasoning Pipelines: Passing JSON flight structures, research insights, and hotel/dining matrices downstream into a synthesis engine for unified schedule generation.
• Core Components of Agentic AI: Breakdown of agent runtimes, instruction personas, tool invocation boundaries, and context memory pipelines.
• User Interface & Export: Building interactive Streamlit dashboards featuring live flight card components and direct .txt itinerary downloads.

Who should attend?

• AI Engineers, Developers, and Solution Architects
• Product Managers and Tech Leaders exploring Agentic AI workflows
• Generative AI enthusiasts interested in multi-agent orchestration
• Software Engineers working with Agno, Streamlit, SerpAPI, and Gemini models
• Anyone looking to build practical, real-world AI applications with live API tool integration

Whether you are building your first multi-agent workflow or looking to enhance your Generative AI applications with real-time tool augmentation, this live session will provide practical reference architectures, design patterns, and code implementations you can apply immediately.