LucidLink – The Real AI Opportunity for Media Is Infrastructure, Not Applications

Peter Thompson, Co-founder & CEO, LucidLink

In nearly every customer conversation we’re having right now, the same shift keeps coming up. Broadcasters and content owners are being proactive. Rather than waiting for AI vendors to hand them a finished product they’re building their own AI-driven tools in-house. That means using the same coding assistants and agent frameworks reshaping software development everywhere else, and going to the market for best-of-breed AI capability rather than another application with its own login and its own way of doing things.

Behind a public face that’s understandably cautious, there’s been fast, technical work happening for a while now. Broadcast technology leaders describe being given real room by management to get hands-on with AI, some with dedicated time most weeks just to experiment and figure out where it fits. If content owners are assembling the workflow themselves, they need best-of-breed AI capability they can pull directly into what they’re already building, without being handed another interface to work around.

MovieLabs made a related point in a recent piece situating AI within its broader 2030 Vision for the industry. Their argument is that AI can only be adopted with confidence if the industry has the right metadata, provenance, and identity foundations in place alongside everything else that vision already covers. If you’re assembling your own AI-driven workflow, what a tool is allowed to see and touch matters as much as what it can do, and it should stay out of the user’s way while remaining fully visible and controllable to the organization deploying it.

Underneath all of this is an infrastructure question: how do you give people, and increasingly agents, access to the content and data they need, without moving it, breaking it, or losing control of the workflow to whichever vendor’s interface gets there first? LucidLink has spent a decade solving this problem for human teams working across locations, clouds, and formats. That same infrastructure works just as well for agentic AI. Our MCP Server extends the same shared filespace to any AI agent or coding tool a team already uses, so agents can read, write, search, and lock files the way a person would, with the encryption, permissions, and audit trail an enterprise already requires. The data stays where it is.

We’re seeing this play out in production, even if still early. One customer uses our MCP server as a lightweight API layer to build internal tools that interrogate their own file estate: how many files exist, who has access to what, when something was last touched, turning metadata most companies barely track into something an agent can act on. Another, outside media entirely, an engineering and construction firm, keeps every project plan and client proposal in one LucidLink filespace feeding a chatbot that surfaces the regulatory, environmental, and structural considerations relevant to a new project, drawn from every comparable project that came before it.

The industry has largely settled on cloud object storage as good enough for the data itself. The opportunity is adding value on top of the storage customers already have, not asking them to move it again. That’s why LucidLink works with S3-backed storage rather than against it, so we can focus on the layer that matters to an AI-driven workflow: access, security, and speed.

Building your own AI-enabled workflow is by no means the easier path. But the industry is taking it, because the risk of standing still is steeper than the risk of not building at all. We’ll be talking about all of this at IBC, in Hall 7 on stand 7.C18, not as a special AI demo bolted onto the side of the show floor but as part of how we show up, because it’s already part of how our customers are working. Come find us.