Ian Wong, Co-Founder and CEO
Fanatics operates a business where demand can move quickly and at a very fine level of detail. A championship run, a player trade, or a new product launch can change what fans want almost overnight, with implications for inventory, pricing, and marketing.
The challenge is not a shortage of data. It is bringing the right information together while there is still time to act. That has been the focus of Fanatics’ work with Summation: first automating recurring analytical work, then bringing more context to live decisions, and now redesigning workflows around AI.
From individual AI to shared context
General-purpose AI can make an individual analyst faster. The harder problem at Fanatics was making that intelligence reusable across a company. One person’s AI session contains their accumulated context; another person’s often starts again from the beginning. For AI to become part of how an organization operates, teams need to be able to work from the same understanding of the business and continue after the original conversation ends.
Fanatics and Summation started by bringing finance, merchandising, and other operating data into a shared governed environment. The work was not just connecting data sources; it was making highly disaggregated data usable together, with shared definitions, business logic, and permissions across systems. Today, that foundation brings together data from across multiple operating teams.
Automating recurring work
Fanatics teams produce a range of analyses, forecasts, and reports, and much of the preparation involves repeatedly gathering the same underlying information.
Sam Turner, Director of Planning, saw the potential when she asked Summation to reproduce a daily tracker her team normally assembled manually. The system worked through the relevant data and returned what she described as:
“Pretty much the exact daily tracker that was pulled manually can now be replicated in a minute.”
What mattered more was that the tracker wasn’t a one-time interaction. It could be saved, rerun, and updated as the underlying data changed. Fanatics teams have since created hundreds of workflows and thousands of outputs, turning recurring analysis into reusable work. For another recurring review, Sam estimated that Summation could prepare roughly 90 percent of a two-hour process, leaving the rest for human review and judgment.
Bringing more context to live decisions
Fanatics’ operators make decisions across hundreds of sites, dozens of major leagues and team relationships, products, channels, and inventory positions — variables that can shift quickly. The next step was to use shared data to improve decisions already happening in planning and marketing.
In demand planning, Summation helped planners evaluate forecasts alongside signals that would otherwise sit in separate systems, including inventory, timing, and product data. Summation did not decide how much to buy. It gave planners a broader view of demand and inventory before purchase decisions were locked.
In digital marketing, Fanatics and Summation built an application that evaluates campaign activity and prepares recommendations for marketers to review. The goal is to make bid decisions based on more than marketing metrics alone — for example, whether the product is available, profitable, and worth pushing.
As Tommy Hwang, Sr. Director of AI Strategy and Business Operations at Fanatics, put it:
“A lot of high-value decisions in commerce are not limited by whether the data exists. They are limited by whether the right context reaches the operator at the moment the decision is being made. Our work with Summation is about closing that gap — bringing together the signals, workflows, and human review needed to make better decisions before the window closes.”
The difference from automation is simple: the output is no longer just a faster report. It is a better-informed recommendation for a decision already in motion.
Building around the workflow itself
As the teams worked through more decisions, some analyses exposed a broader opportunity: redesign the process and build an application around it.
In merchandise planning, Fanatics and Summation are developing a shared planning application where planners work directly with buy quantities, receipt timing, and assortment mix against governed company data before orders lock. The application currently supports hundreds of plans that shape high-stakes buy and inventory decisions. Scott Frank, Vice President of Planning, contrasted the work with a more traditional implementation:
“It would have taken us close to a year to implement something. And it would have cost us millions of dollars. Twelve weeks is extraordinary.”
Related work is underway across operations, finance, and inventory.
The data, analysis, business rules, and human approvals are being brought together around a recurring decision.
A decision loop centered around human judgment
Across the Fanatics work, Summation is not removing people from decisions. It is moving them closer to the moments where judgment matters most: setting constraints, reviewing evidence, approving recommendations, and deciding when to act.
As workflows repeat, prior assumptions, approvals, and outcomes can carry into the next cycle. That gives Fanatics a more durable operating loop: AI prepares the work, people make the call, and each decision can inform the next one.
For a business that moves with its fans, the advantage is not just speed. It is acting with greater confidence and control as demand shifts and business variables update in real time.
How close are you to AI outcomes?
AI readiness isn’t just about having the right data. It’s whether your organization can turn that data into trusted, repeatable decisions.
Many teams see their AI efforts stall here. The model can produce an answer, but the business may not be ready to act on it, route it to the right owner, or build it into how decisions get made.
Readiness, 28 of 40
Adoption, 16 of 40