AI Coaching in Esports: GIANTX, iTero, and the Exclusivity Crack Nobody Will Name

In the transcript of Jack Williams' interview about iTero, two section headings sit side by side on the same page. The first covers working exclusively with GIANTX and the likelihood of being copied. The second covers AI-assisted cheating. For three weeks, every time I sat down to write, I opened those two lines again.

The esports world is pouring itself into the second heading. Forums, newsletters, podcasts — all asking the same thing: is AI helping people cheat. That is the easy question, because it arrives with a villain already attached. The first heading is far harder, and because it is hard, it gets skipped. A tooling company signing an exclusive deal with a permanent member of a closed league, where nobody is relegated, carries more weight than every cheating argument combined.

The crack always shows before the collapse; people just prefer the sound of the collapse.

Context: an article whose skeleton is all I can see

Let me be direct about the source. The extract in my hands contains thirteen information points, and ten of them describe the author of the original piece — his dream of one day recreating Natus Vincere's run at gamescom, lifting the Aegis of Champions. Only three points concern the actual subject: Jack Williams, iTero, GIANTX, and the future of AI coaching. Of those three, two come only from headings, not from body text.

I choose to say that plainly, because an analysis built on silence collapses faster than one built on bad data.

What I know for certain: Jack Williams is the interviewee, appearing as a representative of iTero. iTero is a coaching tool that applies artificial intelligence. GIANTX is an EMEA-based esports organisation, formed through the merger of Excel Esports and Giants Gaming, competing inside the European League of Legends system — the LEC, which moved to a franchised model in 2026 with ten permanent teams and no relegation.

One small detail worth noting: the original article says "fourteen years ago" in reference to Natus Vincere winning The International 2026 at gamescom. Simple arithmetic places the piece around 2026. That is the entire temporal anchor I have. No patch, no roster, no pick-ban rates, no schedule. Any claim about the champion meta in this article would be invention, and I decline to invent.

The meta that is actually shifting

Start where few are looking. The meta in this story is not champions, patches, or win rates. It is the toolkit teams use to solve the patch. That meta is moving, and it moves differently depending on the title.

Riot Games ships League of Legends patches roughly every two weeks. At that cadence, the half-life of any learned pattern is short: what you learned last month may be useless this month. The value of an AI tool therefore shifts from "solving the meta" to "detecting the meta skew faster than the opponent." That is a tempo advantage, not a knowledge advantage.

Valve runs Dota 2 the opposite way: major updates are infrequent and system-warping, but between them lie long stretches of stability. There, models trained on historical data stay valid longer. Advantage shifts from tempo to depth.

A product sold to both ecosystems with one identical pitch is a red flag. If iTero tells a Dota 2 team and a League of Legends team that its tool delivers exactly the same thing, then either one side is buying the wrong product, or both are being sold a story. That is the first thing I would check, if I ever sit across from Jack Williams.

An exclusive deal inside a league with no trapdoor

This is the most ignored part, and the heaviest.

The LEC runs as a closed franchise. Ten teams are permanent members. There is no relegation slot, no promotion threat from below. The structural consequence is concrete: an advantage held by one member does not get competed away across seasons. It persists, accumulates, and becomes part of the environment every other team must live inside.

In an open circuit, the story is different. Weak teams drop out, advantages dilute over time, and the market self-corrects. In a closed league, no self-correction mechanism exists. If a tool genuinely changes competitive outcomes, letting it belong to one member is a policy decision, not a commercial one.

League operators will soon have to choose. Either mandate equal access: every team gets the same tool, or at minimum the same data standard. Or restrict tooling: limit what may be brought into the analysis room. Esports history leans toward the second option, and it always arrives late.

I have watched enough seasons to remember how coach regulations tightened step by step. First coaches were allowed inside the match room. Then pushed into the hallway. Then restricted on when they could speak. Every tightening drew a group insisting nothing had changed, and every time, they were wrong. The match truly begins when the whistle blows and the analysis room turns its lights on.

The grey window between games

The AI cheating debate in this article almost certainly concerns pre-match, between-game, and post-match preparation. The reason is simple: real-time assistance during play is unambiguously banned in every major title. There is nothing left to argue about there.

The real grey zone is the window between game one and game two in a BO3 or BO5. In that brief stretch, a tool can aggregate data, suggest draft adjustments, or flag which side of the map the opponent is leaning toward. The line between "analysis" and "direction" is thin enough that on paper it barely exists.

This is where disguise matters. Do not ask what tool a team uses; ask what role that tool is wearing as a costume. The same product, three different coats. The scouting coat: collecting and structuring opponent data. The draft assistant coat: proposing options while humans decide. The crutch coat: the tool makes the decision and a human presses the button.

All three coats sound identical in every press conference, and the behaviour on the coaching bench is entirely different. I have watched enough close-up camera cuts in finals to know that what happens during a between-game break does not match what the team recounts afterwards.

Behind every contract is a silent brain screaming.

The question of being copied, and the better question

The first section of the interview concerns the likelihood of being copied. That is a correct question placed in the wrong spot.

Copying an AI tool is not hard when its edge lies in the model and the data. Models can be rebuilt, data can be bought, and the best people in this industry habitually change employers. That edge evaporates within a few seasons.

Copying is far harder when the edge lies in the feedback loop between the tool and the coaching staff: how a team asks its questions, how it doubts the output, how it converts an index into a specific draft decision at minute twenty. That does not ship inside the product box, and it does not sit inside an exclusivity contract either.

Here I return to a subject I have chased for years in football: spending that escapes the audit. Transfer fees get examined under a magnifying glass, while signing-on fees for free agents drift past almost unquestioned. Exclusivity deals for tooling in esports run on the same logic. The cost does not sit on the payroll, violates no spending cap, and appears in no league financial report. It is a competitive advantage purchased through a corridor nobody guards.

And this is where history speaks. When Moneyball entered baseball, the Oakland Athletics' edge lasted exactly until those numbers became common knowledge. After that, the edge died, and every team had an analytics room. Every surprise on the field is an appointment we arrived late to.

Where I could be wrong

My argument stands on fairly thin ground, and I need to say so myself.

I have two section headings and one name. The entire reasoning about exclusive structure is inferred from proper nouns, not from a single quoted statement. If the interview is in fact only about GIANTX using a tool to save analysis time, then everything above is a building erected on nothing.

Then there is the possibility that exclusivity is good in the short run. It subsidises research, keeps small companies alive through the pre-revenue phase, and in many other industries these very deals funded infrastructure the whole sector later shared. If iTero survives on GIANTX and then opens its API to everyone two years later, I have told the wrong story.

And the real crack might lie elsewhere: data ownership. Who owns a player's telemetry — the team, the player, the publisher, or the tooling vendor. That is a larger question than exclusivity, and it does not appear in either heading I have.

I hold this position until new data arrives. Changing your mind when data arrives is not weakness; it is a different kind of provocation.

What I will be watching

I will place a bet on a checkable marker. Within eighteen months, either the LEC publishes an equal-access clause for third-party analytics tools, or a mid-table team signs an exclusive deal with a rival vendor and forces the league to rule. One of those happens, and it happens quietly, in a four-line statement.

The earliest signal to watch is not an exposed cheating scandal. It is the arrival of a tooling registry — a public list stating which team uses which tool and which data enters the analysis room. The moment a league publishes that list, we will know the crack has been seen. Until then, the collapse has not arrived.