Jack Williams, iTero and GIANTX: The Real AI Fight in Esports Lives in the Between-Game Window

At a fanless arena in Berlin, between game two and game three of a best-of-five, there is a laptop that belongs to no one. It belongs to a model. For roughly seven minutes — shorter than a football half-time, longer than a water break — the team's coach opens it, reads a probability table, crosses two names off the projected draft, and closes it. Nobody in the arena knows. Nobody on the broadcast mentions it. But game three was decided in that exact moment, before the referee pressed the button to start. That is the window Jack Williams is talking about when he answers questions about iTero, about GIANTX, and about the future of AI coaching in esports. Context first, because this story only reads properly once you know who is sitting at each end of the table. iTero sits in the category of analytics and coaching-support tooling — vendors selling answers rather than selling players. It is a market with an odd structure. It does not sell tickets to fans. It does not generate broadcast revenue. Its customers are a small set of organisations inside a single league, and every contract is a bet on one client. The death of one client is the death of one revenue line. GIANTX is the more familiar name. The organisation was formed through the merger of EXCEL ESPORTS and Giants Gaming, announced in late 2026, and entered the LEC — Riot Games' League of Legends EMEA Championship — from the 2026 season. A young brand, sitting inside a closed league, signing an exclusive tooling agreement with a single vendor. That is the context, and it is also the problem. A date anchor is needed to place the conversation. The piece references Natus Vincere lifting the Aegis of Champions at Gamescom "14 years ago." Na'Vi won the first The International at Gamescom in Cologne in August 2026. That subtraction places the article at roughly 2026. It is also a narrow B2B thought-leadership piece rather than mainstream esports reporting — which explains why it carries no bracket, no schedule, and not a single operating number to cross-check. The interview is framed by two headings. The first covers working with GIANTX exclusively and the likelihood of being copied. The second covers AI-assisted cheating. One commercial frame, one integrity frame. Between those two frames sits a gap the conversation never touches, and in my reading, that gap is where the match is actually being played. Based on my experience following matches across several European seasons, I have consistently found that debates about tooling get steered off course by the way the tooling is headlined. Call a tool an "assistant" and people argue about ethics. Call it a "product" and people argue about price. iTero is put in both sentences at once, so the argument never lands in the right place. The commercial frame first. An exclusivity deal with one LEC organisation means that for a period, one team in the league gets the tool. The question Williams raises — the likelihood of being copied — is the vendor's question: which moat holds the advantage, and how deep is it. In analytics software, the moat rarely sits in the algorithm. Algorithms can be copied inside a season. The moat sits in exclusive data and in contractual relationships. The integrity frame comes next. AI-assisted cheating is a topic with real substance, but it is narrower than it looks. In-game, real-time assistance is already unambiguously prohibited in every major title. There is nothing left to debate there, because the rule was written long ago. If a heading about cheating still needs a long conversation, then the conversation must be about the remaining grey zone: the window between games in a best-of-three or best-of-five. And this is where I want to pin the image from the opening. Seven minutes between game two and game three is the only grey zone left open. It is open because it does not exist inside the game — no server logs it, no file records it, no referee sees it. It exists in one room, on one laptop, in the eyes of a coach who has seven minutes to decide a season. The third frame has no heading at all. It is the league-fairness frame. An exclusivity deal in the LEC does not create unfairness. It exposes the unfairness the franchise model already built. The LEC is a closed league: no relegation, no promotion. In an open system, a last-place season is punished with removal from the league; structural advantages get competed away over time because weak teams disappear from the map. In a closed system, structural advantages do not get competed away. They stay. They compound. They become part of the brand, and the brand gets sold back to sponsors as a story. That is why an exclusive tooling contract carries different weight in the LEC than in an open qualifier. In an open qualifier, the team without the tool loses and is eliminated. In a closed league, the team without the tool loses and is still there — losing for another season, and then another, until the gap becomes an identity. There is a precedent league operators already know well. Coach access to players during the pick-and-ban phase was progressively tightened over years, across multiple titles, from free communication to time limits to physical presence restrictions in the match room. That regulatory gravity never disappears. It is simply waiting for a tool strong enough to pull it back into motion. Then comes the variable the interview leaves entirely blank, and analytically it is the biggest gap of all: patch cadence. Valve and Riot operate on two opposing philosophies, and those philosophies produce two different business models for the same category of product. Dota 2 receives infrequent but system-breaking major updates, with long stretches of stability between them. A model trained on historical data retains value across a longer window — which favours deep statistics and historical modelling. League of Legends receives a patch every two weeks. There, the half-life of any learned pattern is compressed, and the tool's value shifts from "solve the meta" to "detect the meta delta faster than the opponent." Those are two different promises. A knowledge advantage in one title is a tempo advantage in the other. A single product marketed identically across both markets is a red flag rather than a selling point. And to judge whether iTero holds durable edge, you would need the target patch cadence, the version-lock rules separating the tournament server from the practice server, and the permitted data-availability windows. The interview supplies none of them. Every performance claim inside it is therefore unverifiable: no sample size, no evaluation methodology, no benchmark. This is where I have to be blunt with the people who are excited about this debate: it is being placed on the wrong axis. The popular framing is "is AI cheating." But that question already has a legal answer in every major title, and once an answer exists, it stops being interesting. The real axis is ownership of preparation infrastructure. Who holds the answer, who rents the answer, who is locked out of the answer. That is a question about ownership, and ownership is far less romantic than cheating. People call it meta; I call it fear, digitised. Here, the tool itself is the digitised fear. A coach trusts the model not necessarily because the model is right, but because trusting himself for seven minutes is scarier. The probability table is not a prophecy. It is an emotional crutch with an interface, and interfaces always sell. And here is the hardest rebuttal: an exclusivity deal does not make the LEC less fair. It only makes the unfairness that already existed visible. Franchised leagues sell permanence to their members, and once permanence is sold, there is no mechanism left to compete away a structural advantage. League operators will soon face two options: mandate equal access, or restrict the tool. Neither is a technical decision. Both are political decisions, and both will be presented as technical ones. The concern is not that tools get stronger. The concern is that the law's adjustment speed is slower than the product's deployment speed. The habit of saying "if it's legal, it's fine" is a dangerous habit in an industry where products ship quarterly and rules ship annually. Law is a lagging indicator. It gets written after the damage, not before. The people currently operating inside the between-game window are not breaking the rules. They are ahead of them. The match begins when the coaching staff submits the lineup, not when the referee blows the whistle. But this season, the lineup is being resubmitted at the seventh minute of a between-game window, on a laptop that belongs to no one. The question for next year is not whether artificial intelligence belongs in the coaching room — it is already in there, sitting next to the coach, and nobody can push it out. The question is who owns the answer sheet, and whether a league that sells permanence to its members can order everyone to read from the same answer sheet.