On March 3, 2026, DraftKings stood in front of investors and described something no major US operator had attempted before: a single app, a single wallet, and a single behavioral data layer spanning Sportsbook, Casino, Predictions, and Jackpocket Lottery. The “Super App” framing was deliberate—this is not a UI refresh or a cross-sell banner. It is a structural infrastructure bet, and it changes the personalization benchmark for every operator in the market.
The timing was equally deliberate. Phase one launched for NCAA March Madness 2026—the highest-volume, highest-stakes week in US sports betting. DraftKings chose to stress-test its unified architecture when the data volume is at its annual peak. That is a signal about confidence, not a coincidence.
This article unpacks what DraftKings actually built, what the AI infrastructure behind it produces in measurable terms, and what the implications are for mid-market operators whose stacks were never designed for cross-vertical personalization.
The LaunchWhat DraftKings Actually Built
The Super App announcement was precise about what changed architecturally. DraftKings became the first major US operator to consolidate all gambling verticals under a single account and unified wallet. A player’s funds, history, and identity are no longer fragmented across products—they exist in one place, accessible to every vertical simultaneously.
The product scope covers four distinct verticals:
- Sportsbook—existing core business, US market leader by revenue
- Casino—live dealer, slots, table games in regulated states
- Predictions—the emerging market for event-outcome contracts, targeting a $10B annual gross revenue opportunity at 10–30% higher adjusted gross margins than sportsbook due to the absence of state gaming taxes
- Jackpocket Lottery—acquired digital lottery platform, lower-friction entry point with distinct acquisition demographics
The DraftKings–ESPN account linking, launched ahead of March Madness 2026, adds an external personalization signal layer. When a player links their ESPN account, DraftKings gains access to declared favorite teams, players, and fantasy rosters—extending behavioral inference beyond app activity into media consumption. This matters because declared preferences are far cleaner personalization signals than inferred ones, and they arrive before the player has placed a single bet.
The addressable market DraftKings is building toward: $55B–$80B in US gross revenue by 2030, according to company projections presented at the March 2026 investor day. The Super App’s cross-sell architecture is designed to capture a disproportionate share of that expansion without proportionally scaling customer acquisition cost.
AI at ScaleHow DraftKings Turned Promotions Into a Machine Learning Problem
The most operationally significant disclosure from DraftKings’ 2025 reporting was not the Super App announcement—it was the promotional spend figure. In 2025, DraftKings AI automated and personalized $400 million in promotional spend through what the company calls “smart promotions.” Approximately 70% of all promotional spending decisions are now made by AI models, with that share expected to continue growing.
To understand what this means operationally: promotional spend at a major sportsbook is not a marketing budget line item. It is the primary lever for retention, reactivation, and cross-sell. Every free bet, deposit match, odds boost, and parlay insurance is a promotional decision. DraftKings has moved the majority of those decisions from manual CRM analyst judgment to machine-driven optimization against individual behavioral profiles.
The messaging infrastructure is Braze with Liquid templating—a combination that enables real-time personalized content across email, push, and in-app channels simultaneously. The performance data from Braze’s platform benchmarks for clients using Liquid personalization establishes the ceiling: +199% mobile push open rate lift and +28% email open rate lift compared to non-personalized messaging baselines.
This is not A/B testing at scale. It is dynamic content optimization across millions of individual behavioral profiles, where the “variant” is unique to each user. The cost structure of running this is fundamentally different from CRM team-driven campaign management—marginal cost per personalized message approaches zero once the model infrastructure is in place.
The commercial context for why this matters: 86% of online gamblers opt out due to irrelevant messaging (Optimove, 2023 Report of Players’ Preferences in iGaming Marketing, n=396). The opt-out rate is not primarily a volume problem—it is a relevance problem. Operators who cannot personalize at the individual level are structurally burning their own database.
The Data AdvantageThe Cold-Start Problem DraftKings No Longer Has
Every operator running separate sportsbook and casino products has experienced the cold-start problem: a new customer in one vertical has no behavioral history in the other. When a sportsbook player first opens the casino tab, the personalization engine has nothing to work with. The first interaction is generic. The first bonus offer is one-size-fits-all. The first few sessions are the highest-risk period for churn, and they are happening without personalization support.
DraftKings’ unified wallet eliminates this structurally. A new sportsbook customer who previously played Jackpocket Lottery arrives with a behavioral profile: game preferences, session timing, typical spend levels, deposit patterns. The sportsbook experience is personalized from the first bet. A casino player navigating to sportsbook during a marquee event arrives with declared team preferences from the ESPN account link. Day-one personalization is not an aspiration—it is the baseline.
The retention implications extend beyond acquisition. Consider a losing parlay: in a siloed stack, that event triggers a standard retention protocol, if anything triggers at all. In DraftKings’ unified architecture, a losing parlay on an NFL game can immediately trigger a personalized casino bonus calibrated to that player’s casino history, or a free lottery credit at a stake level consistent with their Jackpocket behavior. The retention loop closes in minutes, not days. This is not achievable without a single data layer.
Third-party data confirms the commercial impact of solving cold-start. VAIX AI (now part of Sportradar) reports a 12% churn reduction for operators deploying AI personalization across their player base. That number is not from a single campaign—it reflects the compounding effect of consistent relevance at each interaction, including the first ones.
The CAC implication is structural: cross-sell replaces re-acquisition. One acquisition serves all verticals instead of requiring separate marketing spend per product. DraftKings management explicitly framed the Super App in these terms at the March 2026 investor presentation—the unified wallet directly reduces customer acquisition cost per active player across the portfolio.
B2B BenchmarksWhat Third-Party AI Personalization Actually Delivers
DraftKings’ numbers are compelling, but they reflect a first-party AI deployment at a company with 40% engineering productivity gains from AI-assisted development and a dedicated machine learning infrastructure. The more useful question for mid-market operators is: what do third-party B2B AI personalization tools actually deliver at real operator deployments?
VAIX AI (integrated into Sportradar’s platform) provides the most granular public benchmark dataset available for the sector. Across operator deployments, the headline figures are:
| Metric | VAIX / Sportradar Benchmark |
|---|---|
| Bet placement uplift | 20–25% |
| Average bet amount (AI-recommended vs. non-personalized) | +34% |
| Casino sessions (betPARX deployment) | +167% |
| VIP activity (Apostemos deployment) | +108% |
| Total bets (Apostemos deployment) | +18% |
| Churn reduction | 12% |
The betPARX result deserves specific attention: +167% casino sessions is not a marginal improvement in a secondary metric—it is a doubling-plus of cross-vertical engagement at an operator that was running a siloed product before VAIX integration. The mechanism is the same one DraftKings has built natively: behavioral signals from one vertical informing personalization in another.
The gap between DraftKings’ first-party capability and what B2B tools deliver is real but narrowing. DraftKings has the architectural advantage of a unified wallet and years of cross-vertical data accumulation. Third-party tools can close the relevance gap on specific use cases—content recommendations, bonus logic, messaging timing—without requiring a platform rebuild.
The Live Betting Gap54% of Bets Are Live—Most CRM Systems Weren’t Built for That
The personalization gap is most commercially acute in live betting. Live betting now accounts for 54% of total bets placed—it is the majority format by volume, not a niche or a growth area. It is the default behavior of active bettors.
The spend differential is stark: live bettors average $1,583.90 per month versus $846.20 for pre-match bettors—an 87% gap in monthly value from the same customer base, differentiated entirely by behavior type. Live bettors are the highest-value segment in any sportsbook database. They are also the segment that demands real-time, in-play CRM responses that most existing systems cannot deliver.
DraftKings’ unified data layer enables in-play triggers. When a bettor is active during a live match, the system knows their current position, the game state, their historical behavior in similar situations, and their cross-vertical context. A player down on a live parlay can receive a personalized in-play prompt within the session window—a live odds boost on a correlated market, a related prop, or a casino credit timed to a natural break in the action.
Operators running fragmented stacks respond on a delay or not at all. The CRM platform may not receive the live betting signal in real time. The bonus engine may not have access to in-play position data. The result is that the highest-value players receive the least contextually relevant CRM interactions—a structural inversion of where personalization investment should flow.
The 86% opt-out rate for irrelevant messaging is particularly damaging in this context. A live bettor who receives a generic promotional push during an active session is not just unengaged—they are actively irritated. The window for a relevant intervention is measured in minutes, and missing it with an irrelevant message compounds the cost of the failed engagement.
Structural ImplicationsWhy Fragmented Stacks Can’t Copy This
The Super App model has a structural requirement that most mid-market operators cannot meet without significant consolidation work: a single data layer. Cross-vertical personalization is not achievable when sportsbook, casino, and CRM run on separate vendor systems with incompatible data schemas. The player profile that DraftKings can resolve in milliseconds requires, in a fragmented stack, a data engineering project to even define—let alone query in real time.
Mid-market operators typically run three to five point solutions: a sportsbook platform (SBTech, Kambi, Sportradar), a casino aggregator, a CRM platform (Optimove, Braze, Salesforce), and separate bonus/promotion management tooling. Each vendor has its own player ID, its own event taxonomy, and its own data retention policy. Cross-vertical triggers require API integrations between systems that were not designed to talk to each other, maintained by engineering teams under competing roadmap priorities.
DraftKings disclosed a 40% engineering productivity improvement from AI-assisted code reviews, and projects approximately $30M in annual savings from AI absorption of tasks that previously required headcount. These numbers matter not just as operational metrics—they signal that DraftKings is widening the cost-structure gap with operators who have not deployed AI in their engineering function. The benchmark is not static; it is accelerating.
DraftKings CEO Jason Robins described an “AI-first mindset” across the entire business—not confined to marketing or CRM but extending to RFP automation, legal opinion generation, engineering assistance, and customer service (100% chatbot containment rate; case history summarization in 2 seconds versus the previous 45–60 minutes of manual tracking). The Super App is not a product launch happening inside a traditional operating structure. It is a product launch by a company that has already restructured its operations around AI capabilities.
What to Do NowClosing the Gap Without Rebuilding Your Stack
The competitive implication of DraftKings’ Super App is not that mid-market operators need to become DraftKings. The unified wallet architecture required years of product investment and a scale of engineering resource that is not accessible to most operators. The relevant question is narrower: which specific high-value personalization use cases can be addressed with B2B tools deployed on top of existing infrastructure, without a full platform rebuild?
Two priorities stand out based on commercial impact and implementation feasibility.
Priority One: Live Betting CRM Triggers
Live betting represents 54% of bets and the highest-value player segment (87% higher monthly spend). Real-time in-play triggers—personalized prompts timed to game state, player position, and behavioral context—represent the highest revenue-per-interaction opportunity in the CRM stack and currently have the lowest penetration among mid-market operators. B2B AI layers that can consume a live odds feed, a player history API, and a real-time position signal can close a meaningful portion of this gap without requiring a unified wallet architecture.
Priority Two: Cross-Product Bonus Logic
The retention loop—losing parlay triggers casino credit—requires cross-vertical data access that most operators don’t have natively. B2B consolidation layers can address this incrementally by acting as a middleware that reads from multiple vendor APIs and writes bonus triggers back into the CRM platform. This is not the DraftKings architecture, but it produces functionally similar outcomes for the highest-value retention events. The 12% churn reduction documented in VAIX deployments was achieved through exactly this kind of B2B-layer personalization, not through first-party platform consolidation.
The 20–25% bet placement uplift documented across VAIX/Sportradar deployments is achievable via B2B tools on existing stacks. The +199% push open rate benchmark from Braze Liquid personalization is a messaging personalization baseline that any operator running Braze can approach with the right content generation layer. Neither requires a Super App.
The broader strategic context: the global AI sports betting market is valued at approximately $9B in 2024 and projected to reach $28B by 2030 at a 21.1% CAGR. DraftKings’ Super App is both a product launch and a directional signal. The operators who close the personalization gap incrementally via B2B tools over the next 18–24 months are building toward a world where the gap to first-party architectures is manageable. Those who wait for a platform rebuild that never starts are not.
The benchmark is live, at scale, and expanding. March Madness 2026 was the proof of concept. The rest of the year will be the proof of commercial impact. For every operator watching DraftKings’ numbers come in, the question is not whether to act—it is how fast and where to start.
SourcesData Sources & References
- DraftKings Investor Day, March 2026 — Super App launch, $400M AI promo spend, 40% engineering productivity gain, 100% chatbot containment
- DraftKings at BofA Conference, 2025 — ~70% of promotional decisions made by AI models
- Braze: Liquid Personalization benchmarks — +199% push open rate lift, +28% email open rate lift
- DraftKings workforce reduction — ~5% headcount, ~$30M annual savings from AI task absorption
- DraftKings $10B prediction market opportunity — 10–30% higher adjusted gross margins vs. sportsbook
- Super App launch coverage — $55B–$80B US revenue opportunity by 2030
- DraftKings generative AI deployment — 2-second case summarization vs. 45–60 minute manual process
- VAIX / Sportradar AI personalization benchmarks — 20–25% bet placement uplift, +167% casino sessions (betPARX), +34% avg. bet amount, 12% churn reduction
- Optimove: 2023 Report of Players’ Preferences in iGaming Marketing — 86% opt-out rate for irrelevant messaging (n=396)