149 None of this is possible without the unglamorous groundwork of data infrastructure. “You can’t bolt AI onto a mess and expect it to work,” says Crownshaw, adding that the first priority is to eliminate silos. “In media, your gold is where three things meet: content metadata, what the audience is doing, and your rights around all of it. In most places they live in three separate systems that never speak to each other. You need to create a connected data foundation across the systems and data you already have, and tools like Microsoft Fabric or Databricks are built for exactly that.” The second requirement is to make content machine-readable. “All the video and audio is useless to AI unless it’s tagged and structured in the right way,” he says. “The content supply chain is the engine. Content intelligence is the brain. Without that brain, there’s nothing to think with.” The third, and most frequently skipped step, is ensuring good data governance from the outset. “What can the AI touch? What did it do? Why did it do it?” asks Crownshaw. “When you’re pointing AI at your own intellectual property and audience data, you need the right security controls, naming conventions and agentic controls. Governance is what lets you actually go live instead of being stuck in pilots forever. Companies that do the plumbing first are the ones that move faster.” Microsoft’s role in this process, Crownshaw suggests, is to be a deliberately noncompetitive, supportive partner. “We're the platform, not the competition,” he says. “Being a platform provider rather than a media owner helps reduce tension around distribution, audience ownership and advertising monetisation.” The technical stack runs from Microsoft Azure and Fabric to Databricks, Azure AI Foundry and Copilot Studio, with security and identity woven throughout. Increasingly critical is the control layer for agent fleets, provided by Microsoft Agent 365. “Everyone’s about to go from one AI assistant to a whole fleet of them,” he says. “The hard part isn’t building them, it’s running them without turning them into chaos. It’s about knowing who’s allowed to do what, which data they can access, where the human in the loop comes in, and what happens afterwards. That’s where we are genuinely different.” Unlocking value with the partner ecosystem Much of the practical value of data reaches media customers through a rich partner ecosystem that turns Azure capabilities into end-to-end solutions. Databricks sits at the foundation, connecting disparate sources and building the knowledge graphs that power intelligent workflows, while IPV’s Curator and Brahma’s AI platforms use Azure content-understanding models to tag and index content at scale. On the production side, MediaKind has already delivered high-profile live events – including the FIFA Club World Cup – on Azure, combining broadcast infrastructure with AIdriven graphics, automated cuts and quality control. Nvidia’s DGX Cloud, Holoscan and upcoming Vera Rubin capabilities supply the accelerated compute needed for advanced AI and immersive experiences, and WSE Sports converts live footage into personalised highlight reels and vertical clips at speed. Downstream, Adobe and Dynamics 365 bring campaign creation and performance tracking directly into everyday tools such as Outlook and Teams, while SymphonyAI monitors abandonment signals in real time to protect subscriptions. System integrators including PwC, Cognizant and Southworks help organisations connect these pieces and embed the necessary governance.
RkJQdWJsaXNoZXIy NzQ1NTk=