How Our Agentic AI Took Over Our Fantasy Football Draft (and Still Delivered a Campaign Playbook)


When I got the invite to join our agency’s Fantasy Premier League draft at 11am on deadline day, I had one problem: no time to research. Instead of scrolling through endless stats and injury updates, I handed the task to our agentic AI - our digital teammate that doesn’t just follow instructions, but researches, strategises, and makes decisions on the fly. The task? Build me a competitive squad.
For the draft, the AI had not only delivered a full FPL draft team – it had followed the same structured, insight-driven process we use to build digital campaigns for our clients.
Kick-off
At 8 Million Stories, we pride ourselves on building digital growth strategies powered by insight, creativity, and AI. This time, the challenge was unconventional: could the system also handle a Fantasy Premier League draft?
The experiment tested three things: whether the AI could act independently, adapt in real time to shifting data like transfers updates and injuries, and produce a draft list refined and ready for kick-off.
The game plan
We treated the draft like we would any new brief. First came discovery. Instead of audience insights and market research, the AI gathered inputs from pre-season statistics, fixture lists, injury updates, and expert blogs. Within minutes it had absorbed a range of sources and was ready to define its strategy.
Next, we set the criteria. Just as we would when building a campaign, we looked for measurable indicators of success. This time it meant playing minutes, rotation risk, and goal potential. The AI quickly filtered out risky options and non-starters, leaving only viable candidates.
With pre-season and transfer news evolving, the AI continued to refine its rankings. Within an hour, the system had created a complete draft team – balanced, data-driven, and bias-free.
First half highlights
A draft list that might normally take days of reading, thinking and last-minute panics was ready in hours. The AI surfaced outsider picks, low ownership defenders, budget attackers, the kinds of players that often slip under the radar but are good towards the end of the draft. There were no sentimental favourites, no nostalgia picks from “back when they were good in 2019”.
It’s far too early to declare victory. The season has only just begun, and anything can happen, but AI is ready with the next set of transactions for the second game week. What’s already clear though, is how seamlessly the AI translated its campaign workflow into a new context: fast research, clear strategy, real-time adaptation. It’s given me a competitive start and, just as importantly, a cultural moment – AI now has a seat at our strategy table (and, apparently, a license for trash talk).
Match Analysis
What struck us most was that agentic AI is not autopilot. It didn’t just follow instructions; it researched, adapted, and made decisions. The parallels to campaign work were obvious. Discovery, strategy, execution, analysis – the framework was identical, only the data sources changed.
The experiment also reinforced what we believe about AI’s role in marketing. Humans and machines are not in competition. AI brings scale, speed, and rigour; we bring context, creativity, and interpretation. Together, the outcome is stronger than either could deliver alone
Full-time
Yes, this is “just” fantasy football. But even at this early stage, it demonstrated something more valuable: that agentic AI can take on a live brief, adapt in real time, and deliver results that stand up to scrutiny.
Whether that leads to a title-winning season remains to be seen. What’s clear already is how seamlessly the workflow translates: discovery, strategy, iteration, optimisation. The building blocks of a campaign are the same – only the dataset changes.
And if the AI does carry us to the top of the table? Well, that’s just extra bragging rights, but so far so good.
