Roy Hotrabhvanon has been thinking about artificial intelligence since before most people in sports had any idea it was coming.

When he founded PlayerData in 2017, the thesis was straightforward: collect as much real athletic data as possible, from as many teams as possible, because one day the technology would exist to make extraordinary sense of it. The data could not be generated. It could not be fabricated. It had to be earned, session by session, athlete by athlete, across thousands of teams around the world.

That foundation is what makes PlayerData's AI meaningful today. The technology has evolved to help every coach and athlete unlock insights that used to require a dedicated sports scientist, a spreadsheet, and hours of manual work to surface.

What AI Actually Means for Your Program

The word AI gets thrown around so casually that it has almost lost meaning. For coaches, it can feel like a buzzword that belongs to someone else's world: the Premier League performance departments, the professional franchises, the organizations with seven-figure technology budgets.

PlayerData is built on the opposite belief. The same AI capabilities that support the world's most sophisticated technology companies should be available to every coach and athlete, regardless of budget or staff size. And the way to make that possible is to build AI that works for everyone, whether you have a full performance staff behind you or you are the only person in the building.

"We wanted to collect as wide a breadth of data as possible from as many users as possible," Roy explains. "Because we know that data can't be generated. It can't be hallucinated. It has to be collected."

That data foundation, built over nearly a decade across thousands of teams, is what makes PlayerData's AI meaningful rather than cosmetic.

Talk to Your Data

The most visible expression of PlayerData's AI investment is CoachMate AI, an assistant built directly into the platform that lets anyone on the coaching or performance staff ask questions about their athletes in plain language and get immediate, useful answers.

Want to know how a player's top speed has trended over the last five games? Ask. 

Want to understand which athletes are carrying the highest acceleration load heading into the weekend? Ask. 

Want a comparison of two players' workload profiles across the last month? Ask.

"There are these powerful AI models that can take information and distill it very quickly, much quicker than a human could," Roy says. "So we should be able to use that to help the coach."

What looks simple on the surface is doing something genuinely sophisticated underneath. CoachMate AI receives a question in natural language, converts it into a query that interacts with PlayerData's database, retrieves the relevant data, interprets it, and presents it in a clear and usable format, including charts and visualizations where appropriate. The coach or sports scientist asks a question. They get an answer. No manual exports, no wasting 10 minutes sifting through a spreadsheet.

For programs with dedicated sports science staff, this means more time for the high-value analytical work that actually requires human expertise. For programs without that resource, it means access to a level of data insight that would otherwise be out of reach.

‍

‍

Why This Was Always the Plan

This is not a feature that was bolted on when AI became a buzzword. PlayerData made two deliberate architectural decisions in 2017 that made everything that followed possible.

The first was building cloud-first. All athlete data flows into a single centralized system, accessible by the user from anywhere in the world, in real time, by any device. No data stuck on a laptop. No manual uploads. No fragmented records across different platforms.

The second was choosing GraphQL over a traditional REST-based API.  That distinction matters more than it might sound. Most sports technology platforms are built on REST APIs, which work like a fixed order at a restaurant: you ask for a menu item and you get everything that comes with it, whether you need it or not. GraphQL works differently. It lets you request exactly what you need and nothing more, which means faster processing, less noise in the data, and the ability to build AI tools on top of it that can query large amounts of information with precision rather than brute force.

"You can request exactly what you want," Roy explains. "It just returns exactly what you want. That means you can process large amounts of data faster and you don't have to download four spreadsheets and then crunch them together."

When AI technology became broadly available, PlayerData didn’t need to rebuild anything. The infrastructure was already there. The API was already built to handle exactly the kind of precise, high-volume querying that AI models require. 

The Intelligence Beneath the Surface

Beyond the AI assistant, PlayerData has been building something more foundational: its own machine learning models trained on proprietary data.

The clearest example is automatic pitch detection. PlayerData has mapped over 90% of pitches in the UK used with its product. By training a model on satellite imagery, pitch geometry, and the shape of real movement data collected on those pitches, the system can now identify a pitch automatically, even accounting for outdated satellite images or subtle positional offsets in the data.

"We are the tool makers," Roy says, drawing a distinction between building on top of existing AI models and building proprietary ones. "We built our own model. We trained our own tools."

This same approach is being applied in new directions. PlayerData is actively developing models to detect and classify athletic movement signatures using IMU data from the unit itself. And the next frontier for CoachMate AI goes even deeper. Roy's vision is a platform that can look at a player's full historical data profile and help answer questions that previously required significant analytical resources: whether a player has the physical capacity to make a specific run, what a four-week training plan should look like to prepare a player for a specific physical demand, or how to benchmark an individual's output against their own historical norms.

"We are very close to that," Roy says. "We could probably do that now if we put our minds to it."

A Better Tool for Every Staff

Coaches and sports scientists are under more pressure than ever to make smarter decisions with less time. Load management, return to play, performance benchmarking, readiness monitoring. The demands on a modern performance staff have never been greater.

AI doesn’t replace the expertise of a sports scientist or the relationships that make coaching meaningful. But it does free up more time for both. For the performance coach who wants to spend less time pulling reports and more time working directly with athletes. For the head coach who wants to make an informed decision quickly without waiting on an analysis. For the strength and conditioning coach managing multiple teams who needs clarity at a glance.

PlayerData's AI is built for all of you. Technology is no longer the obstacle. The question now is how your program will get the most out of its data?

‍

‍

Interested in learning more about how PlayerData could help your program? Submit your information in the form below.

‍

‍