6 Building autonomous assistants

This chapter covers

Behavior trees for robotics and AI apps
GPT Assistants Playground and creating assistants and actions
Autonomous control of agentic behavior trees
Simulating conversational multi-agent systems via agentic behavior trees
Using back chaining to create behavior trees for complex systems

Now that we’ve covered how actions extend the power/capabilities of agents, we can look at how behavior trees can guide agentic systems. We’ll start by understanding the basics of behavior trees and how they control robotics and AI in games.

We’ll return to agentic actions and examine how actions can be implemented on the OpenAI Assistants platform using the GPT Assistants Playground project. From there, we’ll look at how to build an autonomous agentic behavior tree (ABT) using OpenAI assistants. Then, we’ll move on to understanding the need for controls and guardrails on autonomous agents and using control barrier functions.

In the final section of the chapter, we’ll examine the use of the AgentOps platform to monitor our autonomous behavior-driven agentic systems. This will be an exciting chapter with several challenges. Let’s begin by jumping into the next section, which introduces behavior trees.

6.1 Introducing behavior trees

6.1.1 Understanding behavior tree execution

6.1.2 Deciding on behavior trees

6.1.3 Running behavior trees with Python and py_trees

6.2 Exploring the GPT Assistants Playground

6.2.1 Installing and running the Playground

6.2.2 Using and building custom actions

6.2.3 Installing the assistants database

6.2.4 Getting an assistant to run code locally

6.2.5 Investigating the assistant process through logs

6.3 Introducing agentic behavior trees

6.3.1 Managing assistants with assistants

6.3.2 Building a coding challenge ABT

6.3.3 Conversational AI systems vs. other methods

6.3.4 Posting YouTube videos to X

6.3.5 Required X setup

6.4 Building conversational autonomous multi-agents

6.5 Building ABTs with back chaining

6.6 Exercises