AI Coaching in Esports: iTero, GIANTX and the Governance Gap No One Has Written Rules For

Jack Williams doesn't open with a promise about the future. He talks about the silence between two games — that fifteen-minute window anyone who has sat in a tournament room knows: headsets off, screens dark, and a coach who has to decide what the opponent just changed. That silence is where the story of iTero, GIANTX and AI coaching begins.

In the conversation about the future of AI-assisted coaching in esports, two major themes surface: working exclusively with Giant X and the likelihood of being copied, alongside the question of AI-assisted cheating. At a glance those are two different subjects — one commercial, one about integrity. They sit closer together than people assume, and the gap between them is exactly where the rulebook has not yet reached.

Based on my experience following matches, most arguments about tooling in esports do not start with technology. They start with the schedule.

Context: a closed league and an advantage that is never relegated

The LEC is a closed league. Member teams are not relegated, not replaced between seasons. That sounds like an administrative detail, but it determines how an advantage behaves. In an open circuit, every edge gets eroded: weak teams learn from strong ones, strong teams change, the market flattens itself. In a closed league, a structural advantage — say, private access to an analytics tool — is not worn down. It persists. It compounds. It becomes habit, then becomes the baseline.

GIANTX, rooted in the EMEA ecosystem, is one of the long-standing members of that structure. When an organisation like that signs a private deal with a product like iTero, the question stops being whether the tool is good. The question becomes whether the tool creates a new baseline other teams are not measured against.

Fourteen years ago, a team won the very first The International at Gamescom with no data model in hand. The debate has travelled a long way since, yet nobody has written the rule for its middle ground.

Analysis: value is inversely proportional to the number of users

There is a commercial paradox few people name. The value of an analytics tool is inversely proportional to how many people use it. If every team has iTero, iTero is no longer an advantage — it is a floor. If only one team has it, it becomes the league's problem.

Williams is reported to have discussed the risk of being copied. That is not purely an intellectual-property worry; it is a worry about losing the exclusivity that is the entire commercial value. An AI coaching product sells because it is not universal. The day it becomes universal, it is just software.

But there is a deeper layer that patch cadence decides. League of Legends patches every two weeks; the lifespan of any pattern is short. Dota 2 changes on a sparser, heavier rhythm. A model trained on historical data holds its value longer in Dota 2. In League, the best tool is not the one that knows the most but the one that detects the new thing fastest. The same product, two entirely different — even opposite — value propositions.

The key point sits here: pre-match preparation advantage has never been governed as a fair resource; only in-game behaviour has been legislated.

Step by step, publishers have closed the window of direct intervention: no in-game communication, no real-time assistants, no live data streaming into a player's ear. But the window between games stays open. The BO3, BO5 gap — when teams leave the stage and have fifteen minutes to re-read everything — is where AI coaching lives.

If iTero only operates there, it breaks no rule. It just does faster what a good coach already does by eye. But speed is the whole story. A coach takes twenty minutes to notice the opponent has changed gank direction; a system might take two. In a BO5, that gap is not convenience. It is the result.

Rumours about AI coaching run thicker than my ping during a livestream, yet most have no sample size behind them. No public document covers evaluation methodology, the definition of improvement, or sample size. That is not a small detail — it is the entire foundation of every claim.

The counter-intuitive angle: AI does not solve the meta

It is easy to romanticise AI as an immortal coach that knows everything and never forgets. But AI does not solve the meta. It only shortens the time between the meta changing and you noticing.

That means its true value lies not in knowledge but in speed. And speed is something a publisher can take away with one line in a tournament rulebook. When one team turns a preparation advantage into a private asset, the whole league becomes a ranked game waiting to quit.

There is another blind spot nobody in this conversation touches. Nobody discusses what happens when a publisher decides to tighten the rules. If a tool genuinely affects outcomes, the organiser faces two options: force equal access for all teams, or restrict the tool. Both are scenarios that have already happened in history. In-game communication was once permitted, then progressively tightened, then removed. There is no reason to believe analytics tooling is immune to the same spiral.

I have a bad habit of abandoning projects halfway. My bilingual football–esports dictionary has sat still for eight weeks, like a meta nobody has found a counter for. But there is one thing I do not abandon: sitting down after every time I watch a small advantage turn into structural unfairness. Tool exclusivity smells exactly like that.

Closing

The question is not whether AI belongs in esports. The question is: when a preparation advantage becomes the private asset of one member of a closed league, who writes the rule — the publisher, the organiser, or the team selling the tool?

No one is online on the server anymore, but I still hear keyboards echoing from the empty stands.