Ask an AI agent a business question, and it could turn that question into an audience recommendation in seconds. Behind that experience, however, the data could have arrived through a scheduled flat-file delivery under an agreement negotiated by people. The application may be new, even when the mechanisms supporting it are familiar.
That’s part of what makes the shift towards agentic advertising so interesting. The applications are changing quickly, but much of the infrastructure supporting them already exists.
For agencies and technology companies experimenting with AI, the challenge isn’t simply building a smart interface. It’s making sure the data behind it is accessible, reliable and set up in a way that actually supports the use case.