The Challenge:
Champions: ITN, BBC, Channel 4
Participants: CUEZ, Amira Labs, Highfield-AI, Monks, CuePilot, Shure, EVS, Moments Lab, Google Cloud
The Challenge:
Firing all cylinders on progressive, forward-thinking innovation, this ambitious Accelerator addressed the challenges in live TV studio production workflows, many of which have become fragmented and costly due to legacy systems, evolving technology that requires more flexible and scalable solutions, and the increased demand for multi-platform content.
Inspired by the successful R&D of previous IBC Accelerators, including the 2024 ‘Evolution of the Control Room’, this 2025 project proposed AI-driven production assistants to intelligently integrate into control room workflows and enhance live production through AI-assisted automation.
As frameworks and standards emerge for multi-agent systems, the POC set out to leverage Large Language Models (LLMs) and agentic systems to create intelligent user interfaces (UIs) and enable Agent-to-Agent (A2A) interaction to solve complex tasks across the production stack, eg: navigate run orders, detect errors, pre-empt potential issues, connect NRCS data to GFX, recommend content and more.
This approach used the power of these new agentic systems to redefine AI not just as a tool, but as a next-generation, intelligent UI to enable live operators to react with speed, accuracy and confidence, and push the paradigm for accessibility and intelligently connected software and hardware tools.
Kickstart Pitch:
Goals:
To address these Challenges, this POC achieved a groundbreaking, "media-first" live gallery with a customised, voice-controlled "AI Assistant Director".
Built as a new, open, vendor-agnostic framework, it utilised Google-supported Gemini integrations for ADK, A2A and MCP to demonstrate voice-first interaction and become the ‘orchestrator agent’ that coordinated and supervised all agent workflows.
This enabled the AI Assistant Director to run an end to end live news broadcast demo by integrating with the full cast of bespoke agents custom-built to enable Audio, Automation, Checking, Content Discovery, GFX, Music Automation, Rundown, TX and Video Enhancement.
The POC created a new framework of dedicated, strategic AI Agent assistants that all worked together to reduce complexity, speed up decision-making, maintain quality, protect trust, and understand and act on live TV production needs, with an emphasis on human augmentation rather than replacement and integration.
Key Goals & Achievements:
-
Successful development and integration of a working, custom-built AI agentic system with live TV news workflows
-
A new framework created for intelligent and collaborative agentic integrations with production control technology
-
Demonstrated user interactions with agentic technology to action realtime workflows and tasks
-
Built a system (not a product) that is agnostic to the software tested against
-
Utilised AI as intelligent and collaborative UIs and integrations
-
Created impact that successfully tested an AI agentic technology production that enhanced efficiency, reduced cognitive load, and streamlined workflows within a high-pressure, live news environment.
PoC Results:
At a time when the industry seeks genuinely useful AI integration rather than experimentation, this IBC Accelerator porposed and created a structured approach to develop intelligent production assistants that augment human decision-making while maintaining editorial control. The team successfully demonstrated practical use cases and set new standards for AI-driven live production to ensure that next-generation automation remains adaptive, reliable, and deeply embedded in real-world broadcast operations. Meanwhile, the emergence of technical frameworks for agent to agent interaction during this project’s R&D sprint, as envisioned in the original pitch, also enabled the team to test and demonstrate the potential for this approach to fundamentally revolutionise the industry’s approach to system integration.
PoC Results Showcased at IBC2025:
At IBC 2025, this Accelerator conducted live demonstrations of their working AI agentic system to showcase real time, AI agent to agent workflows and tasks for a live studio gallery. This included live demos both on stage during the Final Showcase Session and at the Accelerator Zone.
What’s Next?
Looking ahead, our Champions aim to further explore R&D for utilising existing and new AI agents to assist with live production workflows for news and live show galleries.