An agentic AI ad pipeline for World Cup 2026 by Bria and Nimble – and a Claude skill to run it yourself
Bria.ai
In summer 2026, FIFA brings the World Cup to MetLife Stadium. Every hotel within an hour of the venue is going to be watching the same things: competitor nightly rates, which fan groups are traveling in, which matches are driving demand into which neighborhoods. The winners won't be the brands with the biggest ad budgets. They'll be the ones whose marketing can respond in hours, minutes, even seconds, instead of weeks.
That's the kind of scenario we've been building toward with agentic workflows – and the pieces are already here. At our recent New York meetup with Nimble and Microsoft, we walked a room full of retailers, developers, and founders through two live demos of exactly this pattern. Microsoft's Sue McMahon framed where agentic AI fits into the broader commerce stack, and we picked up from there with the pipeline.
Scenario 1: Answering a competitor's price cut in seconds
Picture a retailer who wakes up to find some big competitor has quietly undercut their flagship product. Traditionally, by the time marketing spots it, briefs an agency, and ships a response, the moment's gone.
In the demo, Nimble picked up the price change in real time and handed it to Bria's Fibo Edit – our deterministic image generation model, built on structured data so the same brief produces on-brand output every time, not a lottery. A full set of reactive ad variants, generated in seconds.
Scenario 2: Gas stations competing on value as pump prices rise
When everyone's pump prices are climbing, you can't win on price, so you change the game. Nimble watched what nearby competitors were offering: loyalty points, coffee combos, car-wash promos. Fibo Edit turned those gaps into creatives that highlighted your value-add. You're not cheaper. You're the smarter stop.
How it works?
Nimble is a real-time web scraping API. Point it at a competitor's page – a product listing, a promo, a pricing update – and it returns the signal as clean, structured data within seconds of the change going live.
Bria Fibo Edit is our VGL-native (Visual Generative Language) image editing model. It reads the structured signal Nimble produces and renders on-brand creatives with consistency you can ship: same brief in, same on-brand asset out. Reproducible, rights-clear, production-ready.
Put them together: a price drop gets scraped, normalized, and rendered into a full campaign: headline, image, sized variants. All before your marketing team's Slack channel even pings.
Back to the World Cup
Now apply that same pattern to, say, a Hyatt property near MetLife. A competing hotel chain quietly drops their World Cup weekend package at 2pm. By 2:15, Nimble has caught it. By 2:20, you've got three ad variants sized for Instagram, X, and a display banner – leaning into whatever your actual advantage is, not a panicked price match. Multiply that across six weeks of the tournament, with dozens of locations, thousands of price changes, and you've got a marketing operation that's effectively always awake.
The Claude skill
The meetup demo ran on two APIs wired together by hand. After the event, a few people asked the obvious question: "Cool. How do I get started tomorrow?"
So we built a Claude skill that does the wiring for you. It's a single markdown file that walks Claude through API keys, the Bria MCP connection, and your first Nimble-to-Fibo Edit workflow. From there, Claude orchestrates the loop itself: read the scraped signal, decide what's worth reacting to, brief Fibo Edit, output the creative.
Grab the file here, drop it into Claude, and prompt it: “Set up Nimble and Bria for me.”
You'll have a working pipeline by the time your coffee finishes brewing.
Agentic marketing is new ground. Nobody has years of experience here yet, which means the advantage right now belongs to the teams willing to build, break things, and iterate out loud. We'd rather be in that camp than the one waiting for best practices to emerge.
Try the skill. Ship something. Tell us what you built, and if you're around New York for the next meetup, come say hi.
