Jack Williams, iTero and GIANTX: When AI Takes the Coaching Seat, Who Owns the Edge?

Seven minutes. That is the entire budget of time a head coach gets between game three and game four of a BO5. Seven minutes to drink water, steady the players, adjust the plan. In 2026, for a team like GIANTX, those seven minutes are also the window in which a machine-learning model returns a parsed read on the opponent's draft data.

Jack Williams, the man behind iTero — an AI-driven coaching platform — spoke about exactly that window in an interview published in September 2026. Two themes dominate the conversation: iTero working exclusively with GIANTX, and the likelihood of the product being copied. A third theme — AI-assisted cheating — is raised but left open.

My spreadsheet is full of formulas, but the answer is always outside the cell. In this story, the cell sits somewhere other than where the public is looking.

Who GIANTX are, and why that detail decides everything

GIANTX is a composite name drawn from two EMEA organisations: Excel Esports and Giants Gaming. The organisation competes inside Riot Games' LEC ecosystem — a franchised league, meaning member teams cannot be relegated. This is not administrative trivia. It is the variable that determines everything downstream.

iTero is an analytics and coaching platform built on AI. By the industry's common description, it ingests match data — draft history, win rates by champion combination, fight timings, jungler pathing — and uses statistical models to generate recommendations.

Technically, esports teams have used data analytics for years. Every tier-1 team has at least one analyst manually scraping data. iTero's novelty lies in three things: processing speed, degree of automation, and exclusive access.

Three time windows in which a tool like this can affect match outcomes:

First, pre-match — opponent analysis and draft planning. This zone is entirely legal in every major league, and it is where every team already invests.

Second, in-match — real-time assistance. This zone is prohibited almost absolutely. No major league permits a coach to transmit tactical information to players during play.

Third, between games — the BO3 and BO5 window. This is where the rules are most ambiguous, and where the entire argument is detonating.

People ask what I look at before a deal closes. I look at motive, not at price. With iTero, GIANTX's motive is plain: buy an information edge inside a window the rules have not yet defined.

The exclusivity deal: selling an edge, not selling software

iTero does not license its software to everyone. According to the interview, the company works exclusively with GIANTX. Williams speaks plainly about the risk of being copied — meaning he knows his business model can be replicated.

This is a problem I have seen many times in the football transfer market. A club does not buy a player because he is good. It buys him because rivals do not have him. Value lives in the gap, not in absolute ability.

With iTero, that gap has a name: exclusive access. If all ten LEC teams used the same AI model, nobody would hold an edge. GIANTX's investment would become a fixed cost, like rent on a training facility. But if only one team uses it, that spend is pure competitive advantage.

On valuation, I always price by opportunity cost rather than by list price. The right question is not "what does iTero cost". The right question is: if GIANTX did not buy iTero, how much would it cost to build an equivalent analytics department in-house? Two full-time data staff plus one machine-learning engineer plus deployment time — that number is not small, and more importantly, it produces no exclusive gap.

The key point: iTero sells an information edge, and an information edge only has value when it is not shared.

Williams talks about the risk of being copied. The market's reflex is "then let them copy it". Look closer and there are three real barriers.

Barrier one is data. A good model needs clean, correctly labelled data, and most importantly data rivals do not have. If iTero has access to GIANTX's internal training data — scrims, coaching notes, practice schedules — then a rival's clone, even with the same model architecture, still lacks the raw material.

Barrier two is time. In esports, every technical solution has a half-life. A copier is always one season behind the inventor. In an industry where coaching contracts typically run two years, being one season late consumes the entire value of the contract.

Barrier three is relationships. This is where most investors misread the model. A credible report needs three signatures: the assistant coach, the agent, and the person in the kitchen. In esports, "the person in the kitchen" is the analytics department and the team's operations staff. Without that signature, every tool is just software nobody uses.

The governance vacuum: the rules have no name for this

Every major league has rules on permitted devices and software at the venue. Those rules were written to stop real-time cheating. They were not written for a tool that runs off-site, operates in the between-game window, and is paid for by the team itself.

That vacuum splits into three unanswered questions.

One: does a between-game AI tool count as a second coach? If it does, limits on the number of coaches in the room become meaningless. A machine-learning model does not need a licence.

Two: if one team signs an exclusive vendor deal, do the other teams have a right to demand equal access? In a franchised league, this question carries far more weight than in an open circuit. In an open circuit, structural advantages get washed out through qualifiers and relegation. In a franchised league, there is no washout mechanism. Structural advantage persists indefinitely.

Three: if a team uses AI to analyse an opponent's public data, is that legitimate analysis or information exploitation? Where is the line when match data is a public asset but the analytical model is private property?

League operators have not answered these questions. And that silence is not neutral. It defaults to permission.

The pandemic did not kill the transfer market, it only stripped the varnish off rules we camouflaged with FFP. The same mechanism is at work here. Leagues did not create the AI loophole. They simply have not yet written rules for something that has existed for several seasons.

Patch cycles: the variable everyone forgets

There is a technical detail both AI advocates and AI sceptics in esports overlook: each title's update cadence determines the value of an AI tool.

Dota 2 runs on Valve's cadence — large systemic patches arrive irregularly, but when they land they reshape the whole system, with long stability stretches in between. A machine-learning model trained on historical data holds accuracy longer. In Dota 2, AI's advantage is model depth.

League of Legends runs on Riot's cadence — patches every two weeks. The half-life of any tactical pattern is constantly shortened. AI's value shifts from "solving the meta" to "detecting the meta delta faster than opponents". That is a tempo advantage, not a knowledge advantage.

The practical consequence is concrete. If a vendor markets the same product to both ecosystems, that is a red flag. A product optimised for a stable patch cannot simultaneously be optimised for a two-week patch. The two problems have different objective functions.

This is the kind of analysis I grew used to in football transfers. Son Heung-min is the lesson: a player's value shifts when he leaves the comfort zone of the media. With AI coaching, a product's value shifts when it leaves the title it was trained on.

The contrarian view: the story is not in the word "cheating"

Most coverage of AI in esports circles around an ethical question: is using AI cheating. I think that question aims at the wrong target, and aims wrong in a way that keeps the industry from ever solving the problem.

If a tool is legal under current rules, using it is not cheating. The definition of cheating does not depend on how fans feel. It depends on the rulebook of the tournament operator. Cheating is a breach of terms, not a breach of expectation.

The real issue sits in a different frame: competitive parity. The right question is not "is AI cheating". The right question is "should a franchised league allow one member to hold exclusive rights to a tool that directly affects match outcomes".

I call that the blind spot of the official story. The interview's two subheadings — exclusivity and copying — frame the issue as a commercial dispute. The subheading on cheating frames it as an integrity issue. The frame sitting between them, league fairness, never appears.

And that frame will decide the shape of esports over the next three years.

There is a memorable precedent. Coaches were once allowed to stand in the match room and talk to players. Then leagues progressively restricted it, then banned it outright. The reason was not cheating. The reason was parity — a team able to hire a better coach would hold a structural advantage. AI coaching is walking the same road, only faster.

Why the interview is worth reading even without data

Based on my experience tracking matches and transfer deals, public statements from tool founders always lack one thing: verifiable numbers. The Jack Williams interview is no exception. No sample size, no evaluation methodology, no win rate before and after iTero deployment.

That does not make the interview worthless. It makes it a document of intent rather than a document of results. And in esports, a tool vendor's intent is usually an earlier indicator than a team's results by two to three seasons.

There is one historical anchor the interviewer returns to: Natus Vincere lifting the Aegis of Champions at Gamescom in 2026, during the first The International. At that moment, no team had a data analytics department. Fourteen years later, a tier-1 team pays for a machine-learning model to read its opponents. The distance between those two moments is the entire content of this story.

Rumours are the only thing in football that never get flagged offside. In esports, claims about AI tools work the same way. They exist outside every rulebook, until someone decides to draw the box.

What happens next

Three scenarios I consider most plausible.

Scenario one: the LEC and other franchised leagues add clauses requiring disclosure of all analytics tools used in competition, mirroring how traditional sports publish equipment. This does not ban AI. It simply turns AI from a secret edge into a public cost.

Scenario two: a rival team signs with a second AI vendor, and the coaching-tool market explodes like the sports-equipment market. iTero loses exclusivity but gains a much larger market.

Scenario three: operators ban outright any tool generating automated recommendations in the between-game window, forcing AI back into a pre-match analytics role.

None of these three excludes the other two. And none is decided by iTero's technical quality. All are decided by regulatory text.

If you are looking for an early signal to price an esports team's next move, do not read the technology press release. Read the tournament rulebook. Clauses on permitted equipment always arrive first, and they always reprice the entire tool market behind them.

The seven minutes between game three and game four will still exist. The only remaining question is: in those seven minutes, who is allowed to sit beside the coach — a human assistant, a machine-learning model, or both.