Introducing the Team Wagering Heavily on AI Betting Bots

Emerging innovators are striving to develop agents that can actually place bets rather than merely providing tips, yet the industry is facing significant challenges. Tom Fleetham, who previously led business development for a blockchain platform named Zilliqa, was involved with an AI gambling agent called Ava, designed to predict horse race victors. “She provided solid analysis and yielded good results,” he explains. “The difficulty arose when it came to actually placing the bets.”

According to Fleetham, the company struggled to get Ava to effectively place bets using crypto wallets reliably and in a timely manner. “It was a prolonged process,” he recalls. “Eventually, we abandoned it.”

YouTube is filled with guides on creating and managing gambling agents that can place bets for users. However, these offerings do not seem to be producing new millionaires—or even thousandaires. Siraj Raval, a YouTuber who shares tips on making money with AI, has highlighted a side project named WagerGPT on his channel. He asserts that the tool can place bets and charges a subscription fee of $199 per month for its use. “About eight months ago, I added a feature enabling WagerGPT to place bets. It’s been doing that for users consistently since then,” Raval claims. He states that WagerGPT scans over 40 sportsbooks and identifies various factors that humans may overlook. Raval invited WIRED to join a Telegram group filled with WagerGPT users, but the channel was mostly inactive, with most recent messages inquiring about the status of the service. “It’s completely dead,” says Pete Sanchez, a participant. “A waste of money.”

Since AI agents cannot manage traditional bank accounts, most fully automated betting products concentrate on sports gambling platforms and prediction markets that accept cryptocurrency, as many agents can operate crypto wallets. Among the notable projects enabling AI agents to conduct various transactions on humans’ behalf is Coinbase’s AgentKit. It envisions a scenario where agents can carry out diverse financial activities—from booking airline tickets to trading cryptocurrencies and, of course, placing bets on sports events. Lincoln Murr, an AI product manager at Coinbase, mentions that many early applications of AgentKit “were speculative” and acknowledged a lack of notably successful outcomes. “I’m uncertain about how profitable these agents really are,” he admits. The one catching his attention is Sire. This project brands itself as “an agentic sports-betting hedge fund” and operates as a DAO (decentralized autonomous organization). “Those types of projects indicate the direction we’re heading,” Murr states.

Sire, which was previously known as DraiftKing but had to rebrand for legal reasons, is gearing up for a relaunch. Max Sebti, the CEO of its parent organization, Score, says the firm employs a mix of public and private data along with “computer vision technology that monitors the games” to obtain the latest information for determining winning bets. Score’s business model is a complex blend of crypto and gambling. Ultimately, the company intends to enable anyone to transfer USD into a wallet that converts it into a stablecoin. Following that, Sire’s AI agents will aggregate this money with contributions from other users to place bets on decentralized sportsbooks and prediction markets accepting crypto, including Polymarket. Sebti notes that the gains will be redistributed back to the community. He describes the initiative as “a very stable, hedge-fund-like product.” While anyone can contribute funds to a wallet, withdrawing winnings incurs a “performance fee” payable to Sire—though this fee can be lowered if the user purchases the company’s crypto token. The service is set to exit beta this month.