Fantasy football nerds are using AI to get an edge in their leagues this year

This fantasy football season, Aaron VanSledright is letting his bot call the shots.

Ahead of the NFL season, the Chicago-based cloud engineer built a custom AI draft agent that pulls real-time data from ESPN and FantasyPros, factoring in last-minute intel like injuries and roster cuts.

Using his background in coding and cloud computing, VanSledright spun up the agent in just a week with Anthropic’s Claude large language models. He also tapped Amazon Web Services tools, including the new Strands SDK, which helps developers launch agents with just a few lines of code.

“Let’s see how well the AI performs against other humans, because nobody else in my league is doing this,” he tells Fast Company.

In this Premium story, subscribers will learn: 

• How players are using AI to improve player selection and create balanced rosters
• The off-the-shelf AI products that anyone can play with
• Why even early adopters are leaving room for human decisions

VanSledright customized the bot for the quirks of his two-quarterback league (where each team must start two QBs) by designing a weighting system that values positions differently depending on how the draft unfolds: Stockpile skill players first, grab quarterbacks next, and then fill out the rest. The bot will stay active all season, offering weekly lineup projections. He hopes it will ease the pressure—and the visceral angst—of making his own picks.

Admittedly “terrible” at fantasy football, VanSledright wanted to see what would happen if he stripped the human emotion out of this season. Could AI outperform his own mind, gut, and memory?

“[League-mates] are going to show up with their emotion-based picks, maybe some facts based on projections and everything,” VanSledright says. “But it’s kind of all who you feel week to week.”

He’s not alone. As generative AI becomes more accessible, fantasy players are experimenting with DIY tools to optimize their teams—from chatbots that suggest sleepers to scouts that analyze rivals’ drafting habits.

Whether these experiments produce game-winning Hail Marys or just new scapegoats when the season goes south, one thing is clear: Fantasy football players are increasingly willing to let machines take on decisions long dominated by human bias.

“Like a scout sitting next to you”

Chatter about AI in fantasy football is picking up. Mentions of the topic across X, Facebook, Instagram, Bluesky, and Reddit jumped 728% year over year, according to the software firm Sprinklr. (Still, with only 14,000 posts in 2025, the conversation remains relatively niche.)

Ryan Laughlin, a longtime fantasy football fan, is taking a different approach: using AI to track and predict his competition. Combining his own machine learning model with historical data from Yahoo’s API and OpenAI’s LLMs, Laughlin—who works in commerce payments at JPMorgan Chase while pursuing a computer science master’s degree part-time—set out to analyze how his league-mates have drafted in past seasons.

The result is a “draft scout” that generates AI summaries of league players’ tendencies, showing who delays drafting quarterbacks and who prioritizes running backs. As Laughlin explains it: People often forget each manager’s style year to year. His tool is designed to “feel like a scout sitting next to you, trying to help you win your fantasy week.”

To test the concept, Laughlin spent a few hundred dollars on Reddit ads targeting fantasy football subreddits, inviting others to try it and share feedback.

“I’ve gotten actually pretty strong feedback from people who say it resonates with them,” Laughlin says. “Like they read their own profile and say it’s actually kind of true and kind of weirdly accurate. They don’t have objections to it, but people aren’t incentivized to share it because if you share with your league, you’re giving away the competitive advantage.”

AI in fantasy football isn’t entirely new. Back in 2015, IBM Watson rolled out tools that crunched data from Twitter activity, coaching stats, football articles, and more. Today, IBM is in its ninth year of a partnership with ESPN. Its Watsonx platform now powers features like an “AI Weekly Preview” and data-driven categories suggesting which players to add, drop, or trade.

But now those tools are reaching scale. Roughly 13 million fantasy players use IBM features (up from 12 million last year), according to Kameryn Stanhouse, IBM’s vice president of sports and entertainment partnerships. Behind the recommendations: 36 billion data points, from player stats and team performance to news coverage, media sentiment, and injury reports.

“There are the typical stats that you’re getting that everyone has access to,” Stanhouse says. “But also the ability to use that unstructured data and be able to scan the web and understand what reporters are saying. If players are getting a lot of positive media sentiment there, we’re attributing value to those as opposed to others.”

“People treat AI as an Oracle”

Startups are also pushing into fantasy sports from other industries. Sourcetable, originally built for stock traders and hedge funds, now offers its AI-driven spreadsheet platform with features tailored to fantasy football. By pulling data from ESPN, Sleeper, and Yahoo, the tool enables deeper modeling and real-time insights. Founder and CEO Eoin McMillan sees potential to expand into other sports, though he acknowledges some trade-offs.

“Maybe I’m just nostalgic, but I do think we lose some excitement from the game as we move increasingly toward stats management,” McMillan says. “On the other hand, Brady and Belichick were clearly early at being disciplined on this game-manager-mode trend, and the results showed.”

Others are building tools to rethink how lineups come together in real time. Phoenix-based data scientist Ben Jensen, for example, created a “draft optimizer” to help assemble a balanced roster during live drafts—something traditional player projections rarely address. His tool accounts for roster constraints, models hypothetical picks from rival managers, and reflects his personal preferences with tags ranging from “strongly like” to “strongly dislike.” It also simulates whether to take or wait on certain players.

Jensen notes that he used AI mainly for coding, not for decision-making. That distinction, he says, prevents blind trust in a system with its own risks. “Too often people treat AI as an oracle instead, and then get frustrated when it isn’t a magic bullet,” he says. “To whatever extent I inform my strategies with AI, I still am accountable for the outcome.”

https://www.fastcompany.com/91396628/fantasy-football-nerds-using-ai-to-get-an-edge-in-their-leagues?partner=rss&utm_source=rss&utm_medium=feed&utm_campaign=rss+fastcompany&utm_content=rss