AI in broadcast is easy to oversell, which is why the conversation during our recent webinar panel on AI and media workflows was so interesting.
To move past the hype, we asked industry leaders from Graham Media Group, Morgan Murphy Media, Fox Television Stations, and Ross Video what AI is actually changing in media workflows right now.
What made the webinar useful was that nobody tried to make AI sound mystical. The discussion stayed grounded in the same basic idea: take repetitive work off people’s plates, connect the systems that already run the newsroom, and keep humans responsible for the calls that matter.
Here’s a summary of what we learned below. If you want the complete examples, use cases, and audience Q&A, watch the full webinar here:
From experimentation to operations
Michael Newman described Graham Media Group as moving beyond isolated individual use and toward embedding AI in everyday operational systems, while also acknowledging that adoption is still uneven and that extending those gains beyond the earliest adopters is the next challenge.
For broadcast engineers, that changes the conversation. AI is becoming part of the workflow layer, which means access, reliability, permissions, metadata, and support all matter.
Freeing journalists to do journalism
Colin Benedict framed the newsroom case in straightforward terms. Let machines handle search optimization, tagging, clipping, and repackaging. Let journalists spend more time reporting, asking better questions, and improving their work.
AI and Media Workflows in 2026 – An exclusive webcast hosted by TV News Chek
Adoption, experimentation, and practical use cases
Tim Joyce made another useful point: adoption depends on a culture that can admit when something does not work. In broadcasting, “mostly right” can still create real problems. As a result, AI deployments often begin with low-risk, high-value use cases where automation can deliver measurable efficiency gains without affecting critical broadcast operations.
Ross Video’s Jenn Jarvis kept the vendor view practical. AI is useful when it solves a specific production problem. Facial recognition and tracking, captioning, asset tagging, semantic search, and script-aware graphics suggestions all fit because they are closely aligned with the workflow and easy to evaluate.
Integration, shared context, and story metadata
Integration is the next issue. AI agents sound impressive, but they are only useful if the surrounding systems can actually talk to each other. The panel kept coming back to the same operational problem: inconsistent metadata and siloed systems speaking different languages.
Breaking news as the stress test
Breaking news is the obvious stress test. In minutes, one story can turn into live streams, social posts, web updates, clips, graphics, alerts, and archive material. The panel’s point wasn’t just that AI could help pull those pieces back together afterward, but that better orchestration and shared story data could make them easier to track, reuse, and publish across systems from the start.
Trust, guardrails, and governance
No one on the panel argued for removing judgment from the process. In fact, the opposite came through pretty clearly. If broadcasters are going to use more automation, especially in editorial and production workflows, they need to know what happened, when it happened, and why the system made a certain choice. That means keeping a clear record of what content was used, what the AI changed, what a person approved, and where the final decision came from.
It also means putting human checkpoints in the right places, especially when the output could affect what goes on air, what gets published, or how the organization’s intellectual property is handled. That is not red tape. It is the practical foundation for trust. Without it, AI stays in the experiment bucket. With it, teams have a better shot at using automation confidently in real workflows.
Built for breaking news
Move faster, keep consistent, and deliver more compelling stories with a newsroom solution purpose-built for the modern 24/7 news cycle.
Editorial quality, audience insight, and better questions
One of the more interesting threads was that AI was not framed only as a speed tool. Fox is testing it to flag questions of fairness, balance, and language. Morgan Murphy is using it to surface the kinds of follow-up questions an editor or viewer might ask. Analytics mattered too, because most teams now have more audience data than they can realistically interpret on their own.
Where to start, and where to be careful
The caution was just as useful. Digital producer roles will change. Master control is not the first place to experiment with autonomous tools. Start where the risk is low, the work is repetitive, and the payoff is obvious.
For engineers, the message is simple. AI will not succeed because someone wrote a clever prompt. It will succeed when the plumbing is right: clean metadata, open interfaces, useful logs, sensible permissions, and systems people can trust when the show is live. That was the clearest takeaway from the webinar, and it may be the most important one.