> Markdown version of [/videos/2017-ai-agents-agentic-ai](https://www.wearedevelopers.com/videos/2017-ai-agents-agentic-ai). Every page supports `.md` or `Accept: text/markdown`. Links point to the HTML versions so they work for humans too. Agent guide: [/agents.md](https://www.wearedevelopers.com/agents.md).

---

# AI Agents & Agentic AI

Reactive AI is obsolete. Learn to build, evaluate, and deploy proactive multi-agent workflows using Python, Google ADK, and the Model Context Protocol.

- **Speakers:** [Xavier Sala Presas](https://www.wearedevelopers.com/@xavier-sala-presas)
- **Event:** World Congress 2026 Europe - Virtual Stage
- **Published:** July 2, 2026
- **Duration:** 26:53
- **URL:** https://www.wearedevelopers.com/videos/2017-ai-agents-agentic-ai

## Summary

The shift from reactive generative AI to proactive, agentic AI marks a critical evolution in how software achieves autonomous problem-solving. This presentation explores the foundation of AI agents—applications capable of perceiving their environment, reasoning, and executing tasks on a user's behalf through external tools. While traditional robotic process automation (RPA) and standard chatbots rely on rigid scripts, true AI agents leverage large language models (LLMs) alongside API integrations to fulfill complex, unstructured intents. Using Google ADK and Python, the discussion outlines how to construct these intelligent workflows, emphasizing the transition from monolithic systems to modular, specialized multi-agent architectures.

A defining feature of robust agent design is the implementation of advanced architectures like graph-based agents, which introduce predictability by dictating sequential, parallel, or looping workflows rather than relying solely on an LLM for routing. The session also highlights the Model Context Protocol (MCP), a standard that drastically simplifies tool integration by allowing LLMs to dynamically discover and connect to external capabilities without extensive custom code. Additionally, dynamic skills management enables specific functions to load only when required, safely preserving valuable context window limits.

Operationalizing these autonomous systems demands a new approach to the software lifecycle. Because LLMs are inherently non-deterministic, conventional unit testing is inadequate. Developers must conduct qualitative evaluations of both the final output and the agent's trajectory—the specific sequence of tools and sub-agents utilized to reach a conclusion. While deploying multi-agent frameworks to environments like Google Cloud Run or Kubernetes Engine has become highly streamlined, engineering teams must continue to navigate latency challenges, runtime costs, and the strict implementation of AI guardrails to ensure ethical explainability and safe execution.

**Keywords:** ai agent development, agentic ai frameworks, google ADK implementation, RAG systems integration, multi-agent architectures, graph-based agent workflows, MCP tool integration, dynamic skill loading, non-deterministic ai testing, agent trajectory evaluation, sub-agent collaboration, serverless agent deployment, GKE deployment, ai guardrails implementation, ethical ai explainability

## Chapters

1. **Comparing AI agents with generative models and chatbots** (00:00) — Highlighting limitations in reactive AI systems demonstrates the need for autonomous agentic capabilities.
1. **Defining autonomous AI agents and their practical applications** (03:22) — Combining reasoning with external tools enables software to execute multi-step user intents autonomously.
1. **Configuring base agent properties using google ADK** (04:51) — Defining models, tools, and separate instruction files shapes precise operational behavior for large language models.
1. **Creating specialized multi-agent systems for complex workflows** (08:35) — Delegating specific tasks to modular sub-agents prevents single massive prompts from becoming overly complicated.
1. **Organizing project folder structures for deployable agents** (10:05) — Structuring files for prompts, tests, and environment variables ensures clean deployment to serverless or containerized environments.
1. **Implementing graph-based and multi-tool agent architectures** (11:33) — Designing explicit execution paths with sequential, loop, and parallel graphs removes reliance on unpredictable language model routing.
1. **Standardizing data exchanges using input and output schemas** (15:45) — Enforcing strict JSON structures and lazy-loading skills optimizes context windows and ensures reliable data handling.
1. **Core components of robust AI agent architectures** (17:46) — Combining external tools, specialized models, and short-term memory forms the foundation for scalable orchestration.
1. **Utilizing agentic protocols for seamless system integration** (18:43) — Adopting standardized frameworks like the model context protocol reduces custom integration code when connecting diverse tools.
1. **Evaluating language models through qualitative execution trajectories** (22:14) — Analyzing intermediate tool calls and routing steps addresses the unreliability of deterministic testing in generative systems.
1. **Deploying AI agents to serverless compute engines** (23:11) — Executing command-line tools deploys autonomous workflows directly into scalable cloud infrastructure and kubernetes clusters.
1. **Managing runtime costs and ethical guardrails in production** (24:33) — Establishing monitoring bounds and explaining autonomous decisions mitigates the inherent risks of unstructured data processing.

## Related Moments

- [Locating AI agents within the broader machine learning landscape](https://www.wearedevelopers.com/videos/1510-on-a-secret-mission-developing-ai-agents) (from "On a Secret Mission: Developing AI Agents")
- [Introduction to building real-world AI agent solutions](https://www.wearedevelopers.com/videos/1538-composable-intelligence-how-henkel-and-microsoft-are-shaping-the-agent-ecosystem) (from "Composable Intelligence: How Henkel and Microsoft Are Shaping the Agent Ecosystem")
- [Rethinking team structures around AI agent capabilities](https://www.wearedevelopers.com/videos/1539-agentic-devops-how-ai-powered-automation-transforms-software-delivery-on-github-and-azure) (from "Agentic DevOps: How AI-Powered Automation Transforms Software Delivery on GitHub and Azure")
- [Architectural patterns for developing robust generative AI applications](https://www.wearedevelopers.com/videos/1512-building-ai-applications-with-langchain-and-node-js) (from "Building AI Applications with LangChain and Node.js")
- [Shifting security models from passive chatbots to active agents](https://www.wearedevelopers.com/videos/2093-from-shadow-ai-to-secure-intelligence-safe-ai-usage-in-the-enterprise) (from "From Shadow AI to Secure Intelligence: Safe AI Usage in the Enterprise")
- [Empowering automated workflows with agentic AI models](https://www.wearedevelopers.com/videos/1539-agentic-devops-how-ai-powered-automation-transforms-software-delivery-on-github-and-azure) (from "Agentic DevOps: How AI-Powered Automation Transforms Software Delivery on GitHub and Azure")

## Related Articles

- [What is Agentic Programming and Why Should Developers Care?](https://www.wearedevelopers.com/magazine/625-what-is-agentic-programming-and-why-should-developers-care)
- [Graph and AI Trends 2026: Why Is AI Running but Not Yet Delivering?](https://www.wearedevelopers.com/magazine/680-graph-and-ai-trends-2026-why-is-ai-running-but-not-yet-delivering)
- [MLOps And AI Driven Development](https://www.wearedevelopers.com/magazine/82-mlops-and-ai-driven-development)
- [From Prototype to Production: Build AI Agents with This Free 4-Course Learning Path](https://www.wearedevelopers.com/magazine/655-from-prototype-to-production-build-ai-agents-with-this-free-4-course-learning-path)

## Related Jobs

- [AI Software Engineer (Germany)](https://www.wearedevelopers.com/jobs/48317-ai-software-engineer-germany) at **Sunhat**
- [Senior AI Agent Software Engineer (Go, Python) (m/f/x)](https://www.wearedevelopers.com/jobs/48277-senior-ai-agent-software-engineer-go-python-m-f-x) at **Dynatrace**
- [Principal Product Manager, Agent Platform](https://www.wearedevelopers.com/jobs/ext/277541-principal-product-manager-agent-platform) at **GitHub**
- [AI Operations Manager (all genders)](https://www.wearedevelopers.com/jobs/48263-ai-operations-manager-all-genders) at **envelio**
- [Principal Field Architect - AI Agents](https://www.wearedevelopers.com/jobs/ext/1442858-principal-field-architect-ai-agents) at **Twilio**
- [AI Full Stack Engineer](https://www.wearedevelopers.com/jobs/ext/1354435-ai-full-stack-engineer) at **Almedia**