How DraftKings Uses AI to Target Bettors Likely to Lose

Building Machine Learning Models for Promotional Spending

The online sports betting platform tasked data analyst Jayden Butts with testing predictive models designed to isolate which customers would respond to promotional incentives by gambling and losing more money. According to six former employees who worked on these artificial intelligence-driven strategies, the technology analyzed dozens of data points for each gambler. These metrics included daily account balances, typical loss amounts compared to bet sizes, and the likelihood of a user stopping their betting activity.

By strict financial logic, the underlying mathematical framework sought out vulnerable usage patterns. “We are looking for traits and features that we can target that indicate a good investment,” Butts said in interviews detailing the internal projects. Under this operational model, the system flagged participants whose betting behaviors indicated they would absorb bonuses and sustain heavy losses.

The Divergence Between Revenue Optimization and Harm Detection

While loss-maximizing algorithms advanced through rigorous deployment, parallel efforts to implement responsible gaming tools faced internal roadblocks. Four former employees reported that DraftKings stalled or squashed initiatives designed to use similar machine learning capabilities for risk prediction and problem gambling detection.

The company developed a model that would have assigned users risk scores based on live betting activity to flag potential addiction risks. However, internal resistance sidelined the project. According to former staff members, the exact betting records powering promotional targeting may contain clear indicators of developing gambling addiction. This structural divergence left operators open to compliance scrutiny regarding why revenue-generating algorithms shipped while harm-detection safeguards on the same data pipelines stalled.

Corporate Defense and Financial Margins

DraftKings executives defended the company’s use of data science, pointing to tangible impacts on enterprise performance. Data science and analytics improved promotion-driven sportsbook margins by 13 percent in 2025, according to executive disclosures cited in reporting by AI Weekly. The personalization engine successfully directed hundreds of millions of dollars in promotional capital.

In an official statement, DraftKings rejected any implication that its marketing practices are unfair or improperly target customers based on their financial distress. The company stated that promotions are “directed toward customers who demonstrate sustained, engaged use of our platform, not toward customers based on their losses.”

Lori Kalani, DraftKings chief responsible gaming officer, emphasized that the enterprise depends on customers betting within their means for entertainment and fun. While Kalani acknowledged that the platform monitors users for risky behavior, she confirmed that the company declined to deploy risk prediction technology because leadership concluded there was insufficient evidence of its utility.

Regulatory Pressures and Industry Implications

The revelation of loss-maximizing algorithms arrives amid heightened scrutiny from state gaming commissions and legal analysts tracking personalization systems in regulated markets. Legal experts note that operators utilizing customer-lifetime-value machine learning models on high-risk consumer groups face mounting questions from state regulators and plaintiff attorneys.

DraftKings terminated Jayden Butts in late 2024 for performance reasons amid a business downturn that led some internal teams to question the long-term viability of aggressive promotional experiments. Following a brief pause in data science operations, the platform resumed machine learning development. Observers anticipate that regulatory bodies in major betting markets, including New York and Massachusetts, will scrutinize automated targeting methods as pending litigation proceeds through the courts.