Create an AI Player That Talks, Acts, and Uses Game State (Part 2) | Convai Unreal Engine Tutorial

Watch a Convai-powered AI teammate solve the second half of a two-player Stack-O-Bot puzzle in Unreal Engine. In Part 1, we gave the AI teammate environment awareness, live game state, and built-in actions like moving to objects and pressure plates. In Part 2, we extend that same character with custom actions so it can interact with a crate, pick it up, carry it to a destination, and place it where it helps solve the puzzle. This tutorial walks through how to create custom pickup and drop-at actions using Convai Actions in Unreal Engine. We define action parameters, use Actor References for scene objects, create action handlers in Blueprint, connect the actions to Stack-O-Bot’s existing interaction system, and handle success and failure states so the AI character knows whether an action worked or failed. By the end, the AI teammate can understand instructions from the player, move to the crate, pick it up, drop it on the pressure plate, wait for the platform, and help complete the puzzle through natural conversation and coordinated action. This is a step-by-step teaching tutorial for developers building AI characters, AI NPCs, AI teammates, gameplay assistants, training simulations, XR experiences, digital twins, and interactive 3D worlds with Unreal Engine and Convai. Below is the step-by-step overview of what you’ll learn in this video: 1. How to create custom Convai Actions in Unreal Engine 2. How to define pickup and drop-at actions 3. How to add target and destination parameters 4. How to use Actor References for scene objects 5. How to create Convai Action Handlers in Blueprint 6. How to connect AI actions to existing gameplay logic 7. How to handle action success and failure states 8. How to make an AI teammate use objects to solve a puzzle Timestamps 00:00 - Custom actions puzzle demo 01:24 - Part 2 overview: pickup and drop actions 02:08 - Setup improvements from Part 1 04:58 - Create custom Convai Actions 06:40 - Create action handlers in Blueprint 08:26 - Action design: movement, pickup, and failure states 10:38 - Connect pickup to the Stack-O-Bot interaction system 12:03 - Implement the drop-at action 15:25 - Final AI teammate puzzle test 16:49 - Outro Helpful links Watch Part 1 of this tutorial series: https://youtu.be/KcJETVhI7KM Download the Stack-O-Bot tutorial project files, Delta files, and setup guide: https://docs.convai.com/staging/plugins-and-integrations/convai-unreal-engine-plugin/getting-started/stack-o-bot-tutorial-setup Install the Convai Unreal Engine plugin from Fab Marketplace: https://www.fab.com/listings/ba3145af-d2ef-434a-8bc3-f3fa1dfe7d5c Convai Unreal Engine setup tutorial: https://youtu.be/n-UG3nmMeZQ Dynamic Context documentation: https://docs.convai.com/api-docs/plugins-and-integrations/convai-unreal-engine-plugin/features/dynamic-context Stack-O-Bot sample project on Fab Marketplace: https://www.fab.com/listings/b4dfff49-0e7d-4c4b-a6c5-8a0315831c9c Convai MCP for Unreal workflow used to set up the scene: https://youtu.be/dQgOzt8BNIk Sign up at https://convai.com to build AI characters for games, simulations, virtual worlds, XR experiences, training applications, and interactive 3D environments. Have questions or need assistance? Visit the Convai Developer Forum for setup questions, bug reports, and technical support: https://forum.convai.com Follow Convai: LinkedIn: https://www.linkedin.com/company/convaitech/ Reddit: https://www.reddit.com/r/convai/ Twitter: https://twitter.com/convaitech Instagram: https://www.instagram.com/conv.ai/ Facebook: https://www.facebook.com/ConvaiOfficial TikTok: https://www.tiktok.com/@conv.ai ArtStation: https://www.artstation.com/convai #convai #unrealengine #aicharacters #ainpc #aiplayer #gamedev #stackobot #fab #ue5 #convaiactions #customactions #dynamiccontext #gameai