AI Coaching in Esports: Jack Williams, iTero and the Unwritten Governance Line

In the summer of 2026, at the Gamescom trade fair in Cologne, Natus Vincere lifted the Aegis of Champions. Fourteen years later, an interview about artificial intelligence in esports begins with that very memory. Not because Dota 2 is still relevant to the present story, but because the writer of the interview chose to open his own story with a moment from his youth, before entering the actual subject matter: Jack Williams, the iTero platform, the GIANTX organisation, and the larger question of whether AI should be allowed to sit at the coaching table.

I read the interview several times, and what made me pause was not the technology. It was the way a seemingly purely technical topic quickly slid into the terrain of governance. When a tool can change match outcomes, it is no longer software. It becomes a competitive asset, and every competitive asset must answer the question: who gets to own it, who gets excluded from it, and who steps in to arbitrate when two parties disagree.

Every contract begins with a person, before it becomes a number. This story is no different. Before valuation figures, before exclusivity clauses, there was a man named Jack Williams sitting down to tell his story, from building a product to signing a private deal with GIANTX to the fear that someone would copy it.

Context: an interview framed around the wrong centre of gravity

Before analysing anything, one thing must be made clear about the source interview itself. Of the thirteen information points I collected from the original, ten described the biography of the writer rather than the content of the conversation. Only three carried substantive content about the main topic, and two of those were recorded only at headline level, not in the body text. This is a serious methodological problem, and I raise it not to nitpick but to set the boundary for what can honestly be concluded.

This means most of what I can analyse below does not come from the internal details of the interview, but from the very structure of the story it inadvertently reveals. The way an article chooses to tell a story often says more about the industry than what it claims.

On the subject: Jack Williams is tied to iTero, a platform presented in the context of AI-assisted esports coaching. GIANTX appears as an esports organisation that has signed an exclusive agreement to use this tool and worries about being copied. If GIANTX is indeed an organisation operating in the EMEA region's League of Legends ecosystem, then the governing framework for everything in this story is the publisher's rules on third-party software and competitive integrity. That is a grounded assumption but not fully verified in the source material, and I will hold to that level of certainty rather than turn it into absolute fact.

The two most clearly recorded headings of the interview are: working exclusively with GIANTX and the likelihood of being copied, along with the topic of AI-assisted cheating. One commercial frame. One integrity frame. Between those two frames lies a third frame that is entirely absent: the league-fairness frame. And that gap is where I want to dig deeper.

Patch cadence: two games, two markets, two fates for any tool

When discussing AI in coaching, the first thing to understand is that this tool does not operate in a vacuum. It operates on a game that is constantly changing, and the speed of that change determines the value of any machine-learning model.

Dota 2, run by Valve, has an infrequent but highly disruptive cadence of major updates. There are systemic overhauls that upend entire playstyles, followed by long stretches of stability. This characteristic favours statistical and machine-learning tools, because models trained on historical data retain value over longer time windows.

League of Legends, run by Riot Games, updates on a biweekly cadence. Such rapid iteration significantly shortens the half-life of any learned pattern. Here, the value of an AI tool shifts from decoding the meta to detecting the meta delta faster than opponents. That is a tempo advantage, not a knowledge advantage.

These two orientations lead to two different business models. A product marketed identically across both titles is a red flag on the technical side. If iTero truly aims for title-agnostic neutrality, it is trying to sell the same recipe for two entirely different soils — barren land for historical modelling and fertile land for tempo modelling.

Based on my experience watching matches across many Dota 2 and League of Legends seasons, I have observed that the value of analytics tools is always mispriced at launch. People tend to pay for speed, while the real value lies in resistance to change. The more a tool depends on recent data, the faster it loses value after a major patch. In Dota 2, a model may survive for several months. In League of Legends, it can die within two weeks.

This leads to an important conclusion: no patch is mentioned in the source material, and therefore no patch analysis can be honestly executed. Any commentary on champion buffs or nerfs, item changes, or map changes would be a product of imagination, not data. I refuse to do that.

But the Dota 2 reference in the interview is a biographical nostalgia reference, not analytical content. When Natus Vincere is mentioned alongside the Aegis of Champions, that is the writer's personal detail, not a signal about the current competitive landscape. Treating it as competitive data would be a category error. And that category error is precisely how many esports articles inadvertently confuse their readers.

Exclusivity structure: when a tool becomes a lifetime asset

This is where the story becomes interesting structurally. The payment structure is where the soul of a deal resides. But in this case, what is being paid for is not a transfer fee, but access.

If GIANTX has an exclusive agreement to use iTero, then what is being bought and sold is not software. It is an information gap relative to the rest of the league. In an open circuit, such an advantage could be erased if rivals find an equivalent tool or build one in-house. But in a closed league, everything works differently.

Consider the LEC model, where member teams are permanent franchises with no relegation pressure. A structural advantage held by one member persists across seasons rather than being competed away. Exclusivity therefore carries far heavier structural consequences in a franchised system than in an open one. This is the point most esports commentary overlooks, because they look at match results rather than league architecture.

An exclusivity deal also raises a governance question for the publisher and league operator. If a tool materially affects competitive outcomes, the operator will sooner or later face pressure to choose: either mandate equal access for all teams, or restrict the tool itself. This is exactly how in-game coach communication was progressively tightened over time. There is nothing new in the logic. History is simply repeating itself with a different class of tool.

It is not FFP that saves football, but the people who sit down when everything collapses. In esports, with no equivalent mechanism, the question is who will be the first to sit down.

The biggest overlooked issue in this interview is the league-fairness frame. The two recorded headings — exclusivity and copying, and AI-assisted cheating — leave vacant precisely the frame between them. One side is a commercial story: how to protect the advantage of the payer. One side is an integrity story: how to prevent cheating. But the frame in between is the hardest one: whether a preparation advantage bought with money is legitimate in a league that claims fair competition.

The honest answer is that nobody knows, including the insiders. And that very uncertainty about legality is the most notable point of the entire story.

The cheating boundary: the empty window between games

Now comes the thorniest part. When discussing AI and cheating, two entirely different types of assistance must be separated.

The first is real-time assistance, meaning AI intervenes while the match is being played. This is clearly prohibited in every major title. There is nothing to debate, and nothing to write an article about.

The second, and this is the truly interesting grey zone, is pre-match, between-game, and post-match assistance. In a BO3 or BO5 series, there is a rest window between games. During that window, AI analyses data, proposes tactical adjustments, suggests bans and picks. That is the window where the parties argue the most, and also the window with no clear rules.

The interview is almost certainly about the second type, because the first has nothing to discuss. Whether an AI proposal delivered between game two and game three counts as a violation of competitive integrity depends entirely on the definition the tournament operator provides. And when no definition exists, every decision is arbitrary.

The irony is that the more effective the tool, the more necessary the boundary becomes. An ineffective tool no one bothers to regulate. A tool that can change outcomes immediately becomes an object of regulation. This is the paradox of success: the better the product, the sooner it faces the risk of being restricted or forced to share.

In football, when goal-line technology and VAR arrived, a similar debate took place. First was a debate about accuracy. Then a debate about consistency across matches. Finally a debate about the transparency of the explanation mechanism. Esports is walking that exact road, only far faster because product lifecycles are shorter.

After 2026, I do not believe in anything called sustainability — only in the capacity to absorb hits. Tool companies in esports will soon learn that a good product does not protect them from governance. It makes governance target them.

Lessons from those who sat down when the system collapsed

I once spent months rereading documents on financial fair play in football and the financial crises of European clubs. There is one lesson I have carried throughout my career: rules are not born from ethics meetings. Rules are born from crises.

Applied to the AI coaching story, this means there will almost certainly be an incident before there is a law. A team accused of using a tool beyond the line. A match result called into question. A team complaining it was excluded from an advantage without explanation. Only then will the operator truly have to put on paper what previously existed only in verbal debates.

This is not pessimism. It is an honest description of how young governance systems operate. Esports has no shared rulebook for coaching tools, just as football once had no shared rulebook for owner cash flows. The only difference is that esports can shorten that lesson if the parties choose to look at the history of another sport.

But I am not sure they will. Every industry tends to believe it is a special case, that old rules do not apply to it. That belief is precisely why old rules recur.

The contrarian angle: the blind spot lies in the unasked question

The most counterintuitive thing about this story is that it is framed as a technology story, when it is actually a governance story.

When an interview focuses on how to protect a product from being copied and how to ensure it is not abused for cheating, it inadvertently sets a hidden assumption: that the core problem is technical and the solution will also be technical. But the real question is not how to detect cheating, but how to define cheating for a class of tool that has never existed before.

The blind spot is here: when competitive tools are commercialised, tool governance becomes part of competition rather than something standing outside it. Teams with a voice at the operator will have an advantage not only in using the tool but in shaping the rules that govern it. This is the uncomfortable truth the sports finance industry has long known.

A second blind spot: every performance claim about the product in the interview is unverifiable. No data, no sample size, no evaluation methodology is disclosed. This does not mean those claims are false. It only means they are unverified, and in my profession, unverified means unusable as a foundation for any conclusion.

I learned this lesson painfully. Years ago, I wrote a piece on a major deal in which I claimed the entire sum was paid at once. In reality, the deal was split into multiple instalments, with conditions tied to the number of matches played. A colleague caught the error and I was forced to correct it. For a month afterward, I reviewed every tape of press conferences, every draft contract, every comparison table, to find the pattern behind the numbers.

Since then, I never write about a deal or a tool from a single source. Every piece must include an analysis of structure and binding clauses, rather than just a dry figure. Applied to the iTero story: an exclusivity agreement is not a minor detail. It is the entire story, told only halfway.

What cannot yet be concluded, and why that matters

I want to devote a section to making clear what I cannot assess, because honesty about the limits of information matters as much as the analysis itself.

There is no information about patches, versions, balance changes, items, or maps in the source material. There is no information about tournament format, brackets, qualification slots, or schedules. There is no information about team rosters, players, or champion pools. Any claim about these aspects would be fabrication, and I refuse to fabricate under the name of analysis.

This means the interview is a thought-leadership product aimed at a business audience, not a mainstream esports article. It is written for those interested in the tools market, not for fans following a tournament. Understanding a piece's true audience is the first step to not assigning it weight it does not have.

One detail inferable from within the article: if Natus Vincere won the Aegis of Champions at Gamescom in 2026, and the article says "fourteen years ago", then the piece dates to around 2026. This is an arithmetic inference from the article's own wording, with high confidence, but I still mark it as inference rather than fact.

A second inference: the AI cheating debate almost certainly revolves around pre-match, between-game, and post-match analysis, because real-time assistance is already clearly banned in every major title. The genuine grey zone lies in the window between games in BO3 and BO5 series. This is where I believe future regulation will focus, and also where organisations with a tooling advantage will have the most explaining to do.

Why this story reaches beyond esports

There is a larger trend of which the iTero story is merely a small expression. In every sport, competitive advantage is gradually shifting from people to systems. A team does not win only because it has better players, but because it has a better preparation process. When that process is sold as a product, it creates a new layer of inequality that sits outside the audience's field of view.

The audience sees match results. They do not see exclusivity deals. They do not see data-sharing agreements. They do not see non-disclosure clauses on methodology. This entire structural layer is invisible to viewers, yet it determines what they are watching. This is why I always stress that modern sports analysis cannot stop at on-field tactics.

In football, we have seen this with multi-club ownership groups, with owner-related sponsorship deals, with player loan networks. Esports is entering a similar phase, only faster and with fewer protective mechanisms. No independent regulator. No sports arbitration court. No strong players' association. Only publishers, teams, and tool vendors negotiating with each other in the dark.

Dehumanisation starts with how we name a person with data. And in this story there is a similar risk in another form: turning a human coach into a software operator. If AI makes the ban-pick decision, where does the coach's credit lie? If AI makes the tactical adjustment, who bears responsibility when it fails? These questions have no answers yet, and they will shape how teams operate for years to come.

What readers should watch next

If this story continues to develop, there are several signals worth watching.

First, whether the tournament operator issues any official guidance on AI coaching tools. Prolonged silence is often a sign that no one has taken responsibility.

Second, whether other organisations sign similar exclusivity deals. If one organisation can, others will follow. This is the classic domino effect of the transfer market, only here what is transferred is not a player but a tool.

Third, whether a specific incident forces the operator to act. As I said, rules usually come after a crisis, not before.

Fourth, and perhaps most importantly, whether coaches and players are heard in shaping the rules. In every sports governance debate, those directly affected are usually consulted last. If that repeats in esports, we will have a rulebook written by people who never have to live under it.

A final thought

I do not know whether iTero will become an industry standard. I do not know whether the GIANTX deal will open a wave of tool exclusivity. I do not know whether AI coaching will become an obvious part of professional esports or be tightened until it is harmless.

What I do know is that this story is not alone. It is the next chapter of a process that has already happened in every sport that became an industry. When an advantage can be bought, it will be bought. When it is bought, it will be protected. When it is protected, others will demand fairness. And when enough people demand fairness, a rulebook will appear — usually after an incident serious enough has already occurred.

As someone who worked as an assistant editor at twenty-five and once had to correct himself for not understanding payment structures, I look at this story with a certain empathy. Jack Williams is in the position of someone who must believe his product will shape the future. The tournament operator is in the position of someone who does not yet want to write rules. Other organisations are in the position of those waiting to see whether to be envious. And the audience is in the position of not knowing that a debate is unfolding behind what they watch.

The transfer market is a broken mirror; those who look long into it will see themselves. The AI coaching story is such a mirror too. It shows an industry growing up, asking itself what it wants to become, and not yet ready to answer.

What I want to see next is not a new product. I want to see a public conversation between the parties, where operators, teams, coaches, and fans sit at the same table. If that conversation does not happen, it will be replaced by a series of incidents, corrections, and hastily written rules after everything has already broken. And as I learned, details are fate — a deal not properly defined at signing becomes a wound that cannot be healed at the moment of dispute.