Led design of Playbook, the company’s most important 2025 launch — a cross-brand AI betting assistant projected transform betting activity across our platforms, becoming a primary driver of engagement and revenue.
Role
Lead Designer
Timeline
8 months
PM
Brandon Cohen
Brand Designer
Calvin Chardelli
Background
What started as a project to improve a core user journey of the Action app (finding the best betting opportunities as quickly as possible) expanded into a 3 feature AI ecosystem to reduce time-to-bet from the moment of inspiration.
• Playbook Chat: a conversational chat assistant for instant bets, odds and game info, and parlay building.
• Playbook Picks Hub: a discovery feed and dynamic betslip with curated top picks based on all available data (expert analysis, betting systems, projections, signals, etc.)
• Playbook Bot: A social media bot that instantly converts shared betslips into preloaded bets at the sportsbooks with the best odds.
I owned the product design vision and execution across all three entry points. I led the chat UX, designed the component system and brand integration kit, collaborated on the social media bot UX, and partnered with product and brand teams to map multi-phase user journeys from early release through post-launch.
Problem Discovery & Early Exploration
As part of general discovery research for the upcoming roadmap, I ran a cross-functional workshop with product, content, and betting experts. The goal was to surface opportunities to simplify the 4 phase betting process (research, place, sweat, analyze). We consistently heard the need for:
One place to find today’s best bets — aggregated across all tools.
A custom parlay builder — personalized based on user data.
Some interesting challenges I faced in the early exploration of this product were how to make bet building feel deeply customizable and relevant while also simple and frictionless as possible. We were able to integrate bet tracking data and favorited teams and sports gathered during onboarding into the recommender engine. Leaning into the need for “instant” answers, I initially explored a focused single-card betting experience. However, through internal testing and team feedback loops, I found that showing only one bet at a time left users feeling limited.
With so many betting opportunities happening simultaneously, presenting multiple options up front could give users more control and a better chance of finding something that interested them.
Pivot to Playbook: From Feature to Multi-Experience AI Platform
After I presented the concept to e-staff, the vision was expanded. What began as a single “best bets” feed quickly evolved once we aligned with broader business goals and market trends.
Business Strategy
Our affiliate presence on social was dependent on third-party partnerships. By building our own AI-powered social media bot, we could own the customer journey and track conversions more effectively.
Technology opportunity
In parallel with the recommender engine, our AI team was developing an image-recognition tool that could read betslip screenshots. Branding this under the same product umbrella created consistency and visibility.
AI momentum
Leadership wanted a bold, flagship product that showcased our data and technology investments across all brands.
The result: Playbook expanded into three connected experiences - Chat, Picks Hub, and Social Media Bot - all powered by the same recommender engine and image-recognition tech:
By unifying these touch points under the Playbook brand, we positioned it as a scalable, embeddable AI product that could live across all of our brands.
Designing the Details
Chat Experience
Because the user needs (quick odds + picks) were already validated in the core app we were launching this for initially, my focus in chat was on execution and scalability. Instead of relying on plain text, I designed a set of modular response components that made dense information scannable and enabled in-line interactions like odds comparison. By treating these as reusable patterns, I established a component library that can scale beyond betting into future markets, such as fantasy lineup recommendations or draft tools.
Brand Identity
I partnered with our brand designer to create Playbook as a recognizable product identity. We balanced brand recognition with host brand harmony so it would be able to be plugged into multiple contexts seamlessly.
Insight Tags
To really be a hub of all our best data and insights, we built our recommender engine based on a ranked set of “insights”. These insight tags are attached to each pick suggested by Playbook, whether that’s a player trend or an edge percentage from our projections model. When we tested the concept with users, they liked this ability to see the “why” behind the pick and the ability to dive deeper into a particular insight.
The Solution
Here’s the end-to-end solution we delivered, bringing together chat, feed, and personalization in a seamless assistant experience.
Outcome
Impact
We launched Playbook in 2025 across our brands as three experiences: a social bet bot, a personalized feed, and an AI chat assistant.
• The social bot drove a 200% increase in bets in its first week
• It shifted the business from one-time promo-code revenue to ongoing rev-share through sportsbook deeplinks
• Annual bets are forecast to grow nearly 4×, from 32M to 121M
• It contributed to further investment in retention products and first-party data opportunities across the business
Learnings
The social bot and feed took off, but the chat assistant struggled to find traction. Bettors didn’t want an answer handed to them. They wanted to do their own research, just faster. The features that worked removed friction: they jumped users straight to odds comparison and a pre-filled bet slip. Going forward, I’d automate the tedious parts of the journey and leave the decision to the user.
Made by Gwen © 2026