iTero, GIANTX and the Governance Boundary of AI in Esports Coaching

The match ends, but the data remains. The question is who owns it.

In late August 2026, Jack Williams — the figure behind iTero, an AI coaching platform in esports — sat down to talk about GIANTX, about the future of AI in competitive preparation, and about what he called the "likelihood of being copied." It sounds like a technical conversation. But I read it as a governance problem.

The central question is not how strong AI is. The question is: when an exclusive analytics tool is granted to exactly one organisation inside a closed league, under what conditions is the rest of that league competing?

I wrote a blog from a rented room in Nha Trang; now probability takes me everywhere. And at every tournament I set foot in, I try to measure the hardest thing to measure: an advantage that is not on the field.

There is a paradox in how this industry talks about technology. When a new tool appears, people debate it as a question of ethics. When it becomes standard, people forget it was once a question of power. The exclusive deal between iTero and GIANTX sits precisely at that intersection — where ethics, commerce, and governance meet, and where nobody has yet taken responsibility for answering.

Context: when AI moved from the stands into the analysis room

Around 2026, when I started manually logging every metric of the V-League, each match took four hours. I sat in a rented room, pausing video, counting passes, drawing heatmaps by hand. That is how an analyst worked when there were no tools. Every conclusion passed through human hands, and every mistake did too.

Eight years later, the story is different. AI in esports is no longer a coach typing a prompt into a chatbot to ask "what does this opponent like to play." It has become an infrastructure layer: collecting match data, standardising it, modelling opponents, and proposing pick-and-ban options in real time. iTero belongs to the group of companies trying to turn that infrastructure layer into a commercial product.

GIANTX is an organisation from the EMEA system, operating inside a closed league — where member teams are not cycled out, where their slots are guaranteed, and where an advantage gained is not "competed away" by relegation pressure. This is the single most important detail, and the one most online commentary skips.

When a team in an open circuit buys an advantage, that advantage may vanish next season because the team can drop out. When a team in a closed league owns a tool exclusively, that advantage persists across seasons — it is not automatically competed away. That is the difference between an investment and a structural subsidy.

In the conversation, two headings were mentioned: one about working with GIANTX exclusively and the likelihood of being copied, one about whether AI can be used to cheat. Both frames — commercial and integrity — are valid. But a third frame sits directly between them, and it is almost never mentioned: fairness inside a closed league.

I watched matches in the EMEA region throughout the past season, and what I noticed was never the game score. What I noticed was the window between games — the window in which every team must convert data into adjustments. If one team enters that window with a trained model and the other enters it with a handwritten sheet, the match is no longer decided entirely on the map.

Core: three data questions that go unanswered

Let's start with what the conversation did not say, because that is where the truth lives.

First: patch cadence. The Dota 2 publisher ships major updates on a sparse and system-breaking rhythm — one patch can change how the meta operates for months. That means a model trained on Dota 2 historical data retains value across long windows. Statistical tooling holds durable advantage there.

The League of Legends publisher is different. A biweekly patch cadence shortens the half-life of every learned pattern. In that environment, AI's value shifts: it is no longer "solving the meta" but "detecting the meta delta faster than opponents." That is a speed advantage, not a knowledge advantage.

People call me "the numbers guy"; I take that as a compliment. Because this is exactly the point an analyst must separate: the same AI product, if sold identically for both a slow-patch title and a fast-patch title, is a red flag. Its real value must invert between the two environments.

Why does this matter for GIANTX? Because if iTero is a title-agnostic tool, then the claim "we help teams prepare better" is unverifiable without patch cadence, tournament server lock rules, and data-access windows. Nobody in that conversation offered those three numbers. No data, no conclusion. That is my principle, and I make no exception for tools advertised as intelligent.

Second: the likelihood of being copied. Williams talks about GIANTX working with iTero exclusively, and about the product being copied. That is a reasonable commercial worry — but it also exposes an internal contradiction in the industry.

If the tool creates a competitive advantage large enough to worry about copying, then that same advantage, held exclusively by one team, produces competitive inequality. If it is not large enough to create an advantage, nobody bothers to copy it. You cannot simultaneously claim "we create a big difference" and complain "others will imitate us" without admitting the cost: information asymmetry.

And when information asymmetry is institutionalised by an exclusive contract, the question is no longer about iTero. The question is about the league operator.

This is where I draw on my own trade. In sports betting, a sharp player does not win by knowing the result. He wins by knowing what the market has not yet priced. Once enough people know, the edge disappears. The esports market is the same: an AI edge exists only as long as it is not widely available. Granting it exclusively to one team in a closed league artificially extends that scarcity — in essence, a disguised competitive subsidy.

Third: the between-game window. I said above that the window between two games is where I look. Let me be clearer. In a best-of-three or best-of-five series, there is a period in which every team is allowed to adjust. Real-time in-game assistance is already explicitly banned in every major title — so there is nothing to debate there. The interesting grey zone is between games: that is when AI can run, synthesise the data from the previous game, and propose changes for the next.

If a tool can process game-one data faster than a human, it creates an advantage within the series itself — not before the tournament, but during it. This is the thinnest boundary between "preparation tool" and "competitive assistance," and it is where publisher rules have not caught up with the technology.

I have read how major-tournament organisers handled coach-communication rules. Over the years they moved from near-unlimited allowance, to restriction, to an almost complete ban on coaches speaking during play. The logic is simple: if an information channel can change results, the organiser must control it. AI in the between-game window sits squarely on that track.

There is one detail rarely mentioned but worth putting on the table: a historical reference in the conversation itself. Natus Vincere won the Aegis of Champions at Gamescom. That was 2026. If the conversation truly referenced that milestone as "14 years ago," the article anchors to 2026. It is simple arithmetic, but it reminds me that in this industry every claim can be cross-checked with arithmetic. That is why I never write a line without asking myself: where is the number.

And that same Na'Vi reference gives me another hint. It shows a 2026 in which Dota 2 teams had almost no analytics infrastructure at all. Fourteen years later, we are debating whether AI may run inside a between-game window. The pace of infrastructure change is faster than the pace of rule change. That is not a new observation, but it is one this industry has not yet digested.

Contrarian angle: inequality is not in cheating

Here I want to say the opposite of what most commentary is saying.

The crowd looks at this story and sees two things: a company selling an AI product, and a question about cheating. They debate whether AI can help a team cheat. But that is the wrong question. AI cheating, if it exists, will be detected and handled quickly, because it leaves traces in access logs and rules already exist to block it.

The real problem lies elsewhere, and it is far less dramatic: resource inequality inside a closed league.

In an open system, weak teams can be replaced by stronger ones, and a tooling advantage erodes through competition over time. In a closed system, there is no such erosion mechanism. A member team owning an exclusive tool keeps its advantage across seasons, accumulates data, accumulates wins, and makes its product better — which makes the advantage even harder to close.

This is a self-reinforcing loop. The correlation is clear, but I must say it plainly: correlation is not causation. One cannot assert with certainty that tool exclusivity leads to better results, because the team that signs the exclusive deal may already be strong, financially capable, and well-coached. The confounding variable here is very large.

But that is precisely why I hold this position rather than a more dramatic one. You do not need cheating to create inequality. You only need an exclusive contract and a system with no self-levelling mechanism. At the level of probability, I put it at roughly 60 percent that within two to three seasons, a major league will have to confront this question formally — either through equal-access rules or through tool restrictions. I can be wrong. But if I am wrong, I want it to be because I read the data too closely, not because I listened to the crowd too much.

And I want to add one thing about fans. They do not oppose technology. They oppose the feeling that the playing field is no longer level. That emotion is not irrational — it is a variable every organiser should build into its model. I treat fan emotion as a variable to be explained, not noise to be discarded.

An empty field does not need spectators; it needs an analyst willing to look. The field opening here is not on the map, but in the meeting room. And that field has no stands, only contracts.

Takeaway: signals for the next round

Three signals I will watch over the next twelve months.

One, whether closed-league organisers issue any rule on exclusive analytics tools — silence is also an answer, and usually the most expensive one.

Two, whether iTero publishes any verifiable methodology, with clear sample sizes and data windows, or whether it remains only claims. An analytics tool that does not publish its methodology is, statistically speaking, no different from an advertisement.

Three, whether a second organisation signs a similar deal — if so, the market treats this as the new standard, and the debate is over before it began.

The match ends, but the data remains. The only remaining question is who gets to read it first.