For the past two years, the conversation around AI in marketing has been mostly theoretical. What will it be able to do? When will it get good enough? How much will it replace?
That conversation is over.
In the last thirty days alone, a brand ran a fully AI-generated commercial during the NBA Finals. Google put a video generation model directly inside its ads platform. TikTok let brands create broadcast-quality videos from a single product image. Meta built a system that generates and tests four thousand versions of an ad before a human ever reviews one.
AI is not coming to marketing. It is already running campaigns, generating creative, and producing national television spots.
A $2,000 Commercial. Prime Time. NBA Finals.
The most striking example of where things stand came from a company called Kalshi, a prediction markets platform that most marketers had never heard of before this month.
Kalshi aired a fully AI-generated 30-second commercial during Game 3 of the NBA Finals. The spot featured surreal, rapid-cut scenes built to reflect the unpredictable nature of real-world events. It ran in one of the most competitive advertising slots in American television.
-
Production time: roughly two days.
-
Production cost: approximately $2,000.
-
Traditional NBA Finals commercial production cost: seven figures.
The team worked with an AI filmmaker who used large language models to write the script and develop a shot list. Those prompts were fed into Google's Veo 3 video generation model to produce hundreds of short clips. The best ones were stitched together into a finished spot.
This is not a proof of concept. It aired. Millions of people watched it. And it cost less than most brands spend on a single social media post.
The Platforms Are Rebuilding Themselves Around This
What Kalshi did on their own, the major platforms are now building directly into their advertising tools.
Google integrated Veo 3.1 into Google Ads in April. Advertisers can now generate video assets from a text prompt or animate a static image already in their campaign, with no production team, no external tool, and no budget line for creative. The feature sits alongside Nano Banana, Google's AI image generation tool, also built directly into Google Ads. Both are available to any advertiser today.
TikTok went further. The platform integrated Seedance 2.0, ByteDance's flagship video generation model, directly into its Symphony Creative Studio. Brands input a text prompt, a product image, or a reference clip and receive a finished video with synchronized audio in a single pass. No external workflow. No separate tool. Built into TikTok Ads Manager and available globally to all paid TikTok advertisers.
Meta announced AI video creation tools inside Ads Manager that let advertisers apply AI-generated voiceovers to existing video, create UGC-style videos from a product image using AI avatars, and automatically generate Reels ads for every product in a catalog without shooting a single piece of footage. Meta stated its goal is fully automated end-to-end ad creation by the end of 2026.
The infrastructure is being laid at every major platform simultaneously. Creative production as a bottleneck is being systematically removed.
What the Experts Are Actually Saying?
Gartner's 2026 marketing predictions, published after surveying hundreds of senior marketing leaders, describe what is coming with unusual directness.
"AI agents will take over many routine customer engagements, shifting marketing from channel-based execution to fluid, autonomous, agent-driven journeys. This will collapse traditional martech architectures and move marketers into roles focused on supervising intelligent systems rather than running discrete campaigns."
Jim Lecinski, Clinical Professor of Marketing at Northwestern's Kellogg School, predicted that in 2026, mobile AI will become an indispensable shopping companion, with consumers using AI not just to search but to actively advise on, compare, and select purchases in real time.
The shift Gartner describes is not gradual. It is the replacement of discrete, scheduled campaign execution with continuous, agent-driven brand presence. Marketing stops being something teams do and starts being something systems manage.
The human role in that future is not smaller. It is different.
What does this mean for How Marketing Gets Done?
The brands responding well to this shift are making the same move in different ways. They are using AI to eliminate the production bottleneck and investing everything they free up into the thinking that AI still cannot replicate.
A brand that knows exactly what it wants to say, who its customer is, what that customer actually needs, and what makes this brand worth paying attention to, can now express that with a precision and scale that would have been impossible two years ago. The tools are available to anyone. The clarity behind them is not.
AI has made the execution layer of marketing essentially free. The question every marketing team is now sitting with is what they do with that.
The ones who use it to produce more of the same will find themselves with more volume and the same results. The ones who use it to raise the quality of what they actually say, and to be present in more of the places where their buyers are making decisions, are building something that compounds.
