Ahead of the Floyd Mayweather-Manny Pacquiao mega-fight in 2015, longtime Las Vegas bookmaker Art Manteris got a call.
The person calling was a friend, Manteris writes in his book, “The Bookie” (which I quite liked). And this person told the then-Stations Casino sportsbook exec that something was wrong with Pacquiao.
“He’s not right, and he can’t win,” the friend added. “That’s all I can tell you.”
Manteris said he’d got similar calls in the past, but that many were “bogus.” This time, however, he was convinced the information was good. He moved Mayweather’s odds to win to -200. An injured Pacquiao then lost convincingly, and Manteris’ book did well, avoiding a potentially "significant loss."
Mayweather and Pacquiao are now scheduled to fight again in September. Manteris, meanwhile, has retired. If something's up, then, who will get the call? Who will move the number?
Well, if it's an online sportsbook, it might just be a bot. At least at first.
- Major operators like DraftKings and FanDuel are increasingly integrating AI to help with sportsbook pricing, personalization, in-play wagering, and prediction market liquidity.
- With AI-driven algorithms and institutional market makers playing bigger roles, the online sports betting industry appears to be evolving away from human bookmakers.
That's because online sports betting in the U.S. is increasingly becoming a business that embraces and adopts artificial intelligence. That AI usage even extends to the lines and odds that are being set, the creation and pricing of betting markets presented to the bettors bettors. And, for some of the major players, this is crucial to their bottom line.
DraftKings held an investor day Monday that included a lot of talk about AI and how it will apply to their business, such as its planned “Super App.”
DraftKings offers an average of 517 live betting options per game in 2026 compared to 124 in 2022, per investor presentation
— Ryan Butler (@ButlerBets) March 2, 2026
The Boston-based bookmaker is one of the biggest online sportsbook operators in the U.S., and it is laying the groundwork for further expansion of its business, including in the still-nascent space of prediction markets.
DraftKings also rattled off a series of stats to show how AI is making it a more efficient company: 100% of code reviews are now AI-assisted, there's been a 40% improvement in "engineering hour productivity," and so on.
But the AI usage extends to the actual betting part of the business as well, DraftKings noted.
“At DraftKings, AI is not a side initiative,” Zach Maybury, the company’s chief technology officer, said during Monday’s investor event. “It is a company-wide force multiplier. It's accelerating product velocity in the Super App, sharpening pricing and trading, improving discovery and personalization, and driving real operating leverage.”
Maybury mentioned some other "very practical ways" that AI is used by DraftKings, including "prompt-based sportsbook merchandising and content management" and "trading analytics [that are] accelerating trader reviews."
While the lingo is a bit tricky, roughly translated this means DK is asking AI to help it set and present its betting menu, and then to keep an eye on what's being wagered on. Anything concerning could then get flagged to a flesh-and-blood human.
An AI overhaul
Still, the AI-driven shift at DraftKings has led to some flesh-and-blood humans being let go by the company.
A recent note from analysts at Citizens investment bank reported a “restructuring” at DraftKings, which included layoffs. The same note said that DraftKings CEO Jason Robins has a keen interest in implementing AI in the business.
“Internally, the company has adopted AI functions to help write RFPs, help engineers to write code, add in chatbots, and draft legal opinions, all which can save on outsourced labor over time,” Citizens analyst Jordan Bender wrote. “Externally, AI is improving its ability to serve content in a personalized manner. For example, 70% of promotional spending is determined by AI, which should increase over time.”
Bender also wrote that the restructuring “could have been larger or more impactful for the model” if not for DraftKings’ recent foray into prediction markets. Even so, it appears that AI has played a role in that effort as well.
Jeanine Hightower-Sellitto, the manager of the company's predictions business, said on Monday that DraftKings plans to launch its own “market-making division” over the next few months.
In other words, DraftKings will become the other side of trades placed on prediction markets. When you click “yes” to buy an event contract for 55 cents apiece, the DK market-maker could be on the hook for that other 45 cents.
Enjoying Covers content? Add us as a preferred source on your Google account“Because bringing our own liquidity to market is how we will deliver tighter pricing and a more consistent customer experience,” Hightower-Sellitto said during the investor day. “This is also where AI and machine learning compound the advantage: sharper probabilities, faster adjustments, smarter risk and better unit economics at scale.”
Battle bots
DraftKings isn’t alone in pursuing more AI opportunities. Its chief rival and the other half of the U.S. online sports betting duopoly, FanDuel, is interested in applying the technology widely as well.
FanDuel's owner, Flutter Entertainment, says this is crucial to keeping competitive.
“We use machine learning, AI technologies, data science and similar technologies in our products, services and infrastructure, and we are making investments in expanding our AI capabilities, including ongoing deployment and improvement of existing machine learning and AI technologies, as well as developing new product features using AI,” Flutter said in its latest annual report. “If we fail to keep pace with rapidly evolving technological developments in AI, our competitive position and business results may suffer.”
In short, then, the two biggest online sports betting companies in the U.S. see AI as key to their businesses. Those efforts are also part of the broader push for AI adoption across the broader economy, with companies seeing it as a way of reducing costs and increasing efficiency. But will AI twig to insider information about an injured boxer or quarterback fast enough to protect an online sportsbook? Time may tell.
AI likely won’t be wracked by the same kind of ethical quandary that Manteris found himself in after the Mayweather-Pacquiao fight. Ultimately, the whole situation left him concerned about the growing ties between sports and sportsbook operators.
“I had never before had inside information of that magnitude, and it changed my whole outlook,” he writes in his book. “The hypocrisy of being involved in sports while simultaneously taking wagers on it just hit home for me.”
This time it's personal
But the bookmaking business has definitely “evolved,” according to Joe Asher, a longtime gaming executive and CEO of Nevada-based Boomer's Sportsbook.
Online sportsbooks don't necessarily need a flesh-and-blood bookmaker anymore to set their lines and odds; those data feeds can be purchased from a vendor.
In an interview with Covers prior to this year's Super Bowl, Asher particularly pointed to the rise of in-play wagering, which is being “algorithmically driven.”
DraftKings, for instance, suggested in its investor day presentation that more than half of its sportsbook handle now comes from in-play betting. The company also said that its live betting options per game now number more than 500, compared to around 120 in 2022.
So, for the bookmaker (or “trader, as they’re frequently called in the modern vernacular,” Asher said) they’re increasingly required not to set the lines and odds but to monitor them for obvious errors or suspicious movement.
Boomer's Sportsbook is coming soon to Nevada. CEO Joe Asher also had some good quotes today on prediction markets:
— Geoff Zochodne (@GeoffZochodne) July 24, 2025
“Seems as though everybody was asleep, and sports betting was legal all along..."https://t.co/VHSZiPXlFF @Covers
AI is also involved in the “personalization” of online sports betting, Asher said. This could mean remembering that a user keeps betting on a certain team and ensuring that the team's betting markets are front and center when that user logs on.
Still, the actual bookmaking part of the business is under pressure from prediction markets as well. The exchanges are purportedly “peer-to-peer,” but it’s not just Sports Bettor X versus Sports Bettor Y. There may be much more sophisticated parties involved, as DraftKings’ investor day comments suggest.
Asher has been a critic of prediction markets, and he’s critical of them on this point as well.
While there may be instances of an actual person wanting to bet a team on a prediction market having that bet matched by another person, this is the exception, rather than the rule, Asher says. More commonly, he suggested, it’s someone like Susquehanna International Group, one of Kalshi’s institutional market makers, on the other side.
“The idea that this is 'peer-to-peer' betting going on is just a fallacy,” Asher said.
