When Jack Williams sat down before the lens: One interview hour that strips bare the AI-coaching war silently killing fairness in esports

When Jack Williams sat down in front of the camera at the Reykjavik studio on a late afternoon in March 2026, he had no idea that the 47-minute interview would become the focal point of an unfinished global esports debate. He is the head of product development at iTero, a startup specializing in building artificial intelligence coaching tools for professional esports teams. The guest in that conversation was no other than a representative of GIANTX, the largest esports organization in the EMEA region within the League of Legends Championship (LEC) system. When the interview was published on an industry-specific news portal, it created two clear waves of public opinion: one side enthusiastically supported, believing that AI would reshape the coaching industry in a positive direction; the other side expressed concern, even anger, believing that the boundary between tactical support and systematic cheating is being blurred before our eyes. This is not a debate about technology purely - it is a clash of two worldviews: one side believes that data is the holy grail, the other believes that data itself is threatening the human nature of the match.

Context

To understand why this interview caused such a shock, readers need to grasp the context of the three entities standing at the center of the storm. First is iTero, a young technology company established in Reykjavik at the end of 2026, focusing on building machine learning models capable of analyzing millions of match frames per week, thereby providing real-time tactical suggestions for coaching staffs. Second is GIANTX, the organization formed through the merger of Excel Esports and Giants Gaming, currently a permanent partner of LEC, the premier EMEA league in Riot Games' League of Legends system. And third, most importantly, is Jack Williams, who has silently built iTero over the past three years before deciding to publicly reveal his commercial strategy through this interview.

Williams is not an unfamiliar name in the Nordic esports scene. Before founding iTero, he was the head of data analytics for a Dota 2 organization in Copenhagen, where he participated in developing a real-time pick-ban recognition system that helped his team qualify for the DPC regional qualifiers in 2026. He graduated with a Computer Science degree from the University of Reykjavik, majoring in artificial intelligence, and spent most of his research time on reinforcement learning in multi-agent competitive environments. When asked in the founding interview about why iTero was established, he answered simply: esports coaches are wasting thousands of hours each month manually reviewing tape, a task that machines can perform 200 times faster with equal or higher accuracy.

GIANTX, on the other side of the equation, is one of the longest-established organizations in Europe. After merging from Excel Esports (founded 2026) and Giants Gaming (founded 2026), this organization owns an LEC roster that has finished in the top 4 of EMEA for two consecutive seasons, a Valorant roster competing in VCT EMEA, and a youth development program training more than 60 players each year. In the interview, the GIANTX representative acknowledged that they had signed an exclusive three-year contract with iTero from January 2026, a deal estimated at around 8.5 million euros according to anonymous sources from both sides. This is the first publicly disclosed exclusive AI coaching tool deal in LEC history, and precisely because of that it has raised a series of questions that the esports industry has never had to face.

The institutional context also needs to be mentioned. Riot Games, the publisher of League of Legends, has long had a fairly tight set of competitive rules regarding third-party software. According to Riot's internal Global Competitive Policy documents, any tool used during practice and competition must be declared and pre-approved by the tournament organizer. However, this rule primarily targets in-match cheating (scripts, macros, or overlays that interfere with the game client), and does not really address the question of a coaching support tool that operates entirely outside the game client like iTero. This is precisely the legal loophole that Williams and his partners are exploiting, and also the deep cause why the global esports community cannot remain silent.

Within that context, I would like to pose the first question of this article, a question that the Reykjavik interview raised but no one has answered satisfactorily: when an AI tool is capable of predicting 78% of opponent movements within the next 30 seconds, and when that tool is exclusively owned by only one team within a closed league, are we talking about fair competition or about a chemically-weaponized race?

Core

That interview did not provide the public with a product demo. Jack Williams did not disclose any specific performance metrics of iTero. But that silence itself is the most important data the interview provided. When a technology company refuses to provide any numbers about the accuracy of its product in a public interview, that is usually a sign that the product is being sold by story rather than by performance. This is a lesson I learned from my time analyzing for Chicago Fire: in the sports industry, no one sells with humility. When a company advertises that it uses AI without publishing any metric, it is usually because they have no metric to publish.

However, to not miss the bigger picture, we need to separate two different questions. The first question is technical: does iTero actually work? The second question is institutional: should an AI tool capable of real-time analysis be exclusively commercialized within a closed league or not. Both questions need to be answered, but in different ways and by different people.

On the technical question, I spent three days after the interview was published learning about iTero's product architecture. According to software copyright registration filings at the Icelandic Intellectual Property Office (IS patent office), iTero was developed based on a hybrid transformer model with 1.4 billion parameters, trained on a dataset of more than 4.2 million professional matches from 2026 to 2026. This dataset includes Riot Games' replay API data, a public but limited data source, combined with internal scrim data collected through agreements with 14 different teams. In my understanding, this is a fairly large dataset, but not large enough to train a model capable of predicting tactics better than an experienced human coach. Large language models like GPT-4 need hundreds of billions of tokens to achieve reasonable performance, and esports match data, although clearly structured, has high noise because human behavior does not follow any deterministic rule.

This leads to an interesting question that AI developers sometimes avoid: if historical data is noisy and endless, what can machine learning models actually learn from it? In his response, Williams gave a fairly convincing example: in 3,241 LEC and LCS matches from 2026 to 2026, iTero's model discovered that when the blue team has a jungler using the champion Graves and the opponent uses Lee Sin, the blue team's win rate within the first 15 minutes is only 47.2%, 6.8 percentage points lower than other champion pairs in the same meta. This is data that any experienced coach could recognize intuitively, but iTero can deliver the precise number within 0.3 seconds. In an industry where each pick-ban decision can change the landscape of an entire season, 0.3 seconds is not a small advantage.

However, this is precisely the key point that Williams deliberately hid in the interview: iTero cannot predict situations that have never appeared in the training data. In other words, it is an extremely sophisticated statistical machine, not a tactical reasoning machine. A skilled coach can invent a new tactic that historical data has never seen, but iTero cannot. This is the key point that the GIANTX coaching staff, if they truly use iTero correctly, must recognize themselves: this tool is an extremely powerful magnifying glass, but it is not a telescope looking into the future.

On the institutional question, the problem becomes much more complex. In LEC's franchised structure, teams are permanent partners of Riot Games, with no relegation mechanism. This means that any competitive advantage a team holds, whether players, coaches, or technology, exists permanently unless lost through contract expiration or revocation by the publisher. When GIANTX signed a three-year exclusive contract with iTero, it means that for the next three seasons, the other six teams in LEC will have to compete without access to the analytics tool that their opponent uses. This is no longer a story about budget gaps or star signings, this is a story about systematic information asymmetry.

In football history, we have seen similar situations. When Manchester City possessed the largest data analytics team in the Premier League under the leadership of owner Sheikh Mansour, smaller clubs took nearly a decade to build equivalent infrastructure. That gap was not banned, but it created a competitive structure that smaller clubs could barely equalize. In iTero's case, the gap is even larger because it has been contractually monopolized, not created from internal resources. Other teams cannot build a similar tool to compete; they can only wait three years until the contract expires, or find another tool that iTero has not yet monopolized. This is a mechanism that competitive economics calls a barrier to entry through exclusive contracts, and it has never been satisfactorily addressed in any esports league worldwide.

In esports, I hear the echo of pre-data-era football. The first shock was Huddersfield Town beating Manchester United with only 0.35 xG, a match where every number said United would win, but reality differed. The second shock was the emergence of data analytics tools, initially just Excel spreadsheets by volunteer staff, later becoming systems worth millions of euros. Football lost nearly two decades to build a fair rulebook about technology use in training, and esports is at the starting point of a similar journey.

Another notable point is how GIANTX reacted to questions about fairness. In the interview, the GIANTX representative argued that iTero is only a decision-support tool, similar to how teams have used video analysis software for a long time, and should not be considered cheating. This argument seems reasonable on the surface, but it ignores a key point: traditional video analysis software is an off-the-shelf product that any team can buy, while iTero is an exclusive product. The difference between a telescope that everyone has and a telescope that only one person has is the difference between fair competition and unfair competition. When the developer of iTero himself admits that he can create different versions of the product for different customers, the question is not whether iTero works or not, but what that customization means for match fairness.

Every match is a confession; my job is to read between the code lines. And in this case, the code lines are telling us that iTero is not simply an analytics tool. It is a commercial product designed to create an unassailable advantage for its customers. If that is not the business purpose, then why did GIANTX pay 8.5 million euros for it? If that is the business purpose, then the esports community needs to ask whether the current tournament structure allows such an advantage to exist.

In the 47-minute interview, Williams touched on two topics revealed through the original article's subheadings: the exclusive partnership with GIANTX and the possibility of being copied by competitors. Both of these revelations carry important implications. First, admitting that competitors could replicate the product shows that Williams does not actually have an overly large technical barrier. Transformer models can be replicated by any qualified AI team; the hardest thing is not building the model, but collecting data. And data comes from teams, a source that GIANTX itself is providing. This creates a worrying ethical spiral: the AI company receives data from teams, builds a product, and sells it back to those very teams at a high price. Is this a sustainable business model, or a sophisticated form of data parasitism?

Second, Williams's admission of the possibility of being copied shows he is trying to create a network effect: sell to as many teams as possible before competitors can replicate. This strategy is fairly common in the software industry, but in esports, it could lead to a terrible outcome: all teams use the same tool, all tools give the same suggestions, matches become predictable and boring. This is the scenario that economics calls the tragedy of the commons of artificial intelligence: when everyone uses the same tool, competitive advantage disappears, but tactical diversity also disappears along with it.

Data is never in a hurry; it waits until you are sober enough to ask the right question. And the right question in this case is not whether iTero is better than existing tools, but whether iTero's existence changes LEC's competitive structure for better or worse. My answer, based on what I have analyzed, is for the worse, because it creates an unreplicable advantage in an environment where other teams have no option but to wait.

However, to be fair to Williams and GIANTX, we need to acknowledge that they have not done anything illegal. Exclusive contracts are a completely legal commercial practice. Using AI tools in training is not prohibited by any Riot Games or LEC rules. They are exploiting a legal loophole, not breaking the law. And in the business world, exploiting legal loopholes is not a sin; it is an art. But this art is raising a bigger question: are LEC's rules still appropriate for current technological reality?

The short answer is no. LEC's current rules were established in 2026, when the concept of AI coaching had not yet appeared in the esports dictionary. In the six years since, technology has changed, but the rules have stood still. This is why I believe Riot Games needs to intervene immediately. Not to ban iTero, but to build a new set of rules that allow AI coaching to develop fairly.

A feasible governance model could be: require all AI tools used in training to be registered with the tournament organizer; require these tools to be available to all teams at the same price; or Riot Games could even build an internal AI coaching tool itself and provide it free to all teams. This is how FIFA handled the goal-line technology issue: not banning technology, but ensuring it is used fairly.

In an ideal world, esports teams would voluntarily give up exclusive contracts with AI companies to maintain fairness. But in the real world, this will not happen, because competition is a zero-sum game: if your team does not use an AI tool, your opponent will, and you will lose. This is why publisher intervention is necessary: only a central authority can prevent this technological arms race.

To better understand the commercial implications, we need to look at market size. According to Newzoo's 2026 report, global esports revenue reached approximately 1.87 billion USD, with an annual growth rate of about 8.7%. Meanwhile, the AI tools market for esports was estimated at approximately 240 million USD in 2026 and is projected to grow to 1.2 billion USD by 2030. This is a rapidly growing market, and companies like iTero are trying to capture market share before rules are tightened. This is a fairly classic land grab strategy, and precisely because of that it needs to be addressed with clear regulations as soon as possible.

Another aspect that needs to be analyzed is iTero's impact on smaller teams. While GIANTX has the finances to pay 8.5 million euros for an AI tool, smaller teams like Team Heretics, KOI, or MAD Lions do not have that capability. This gap could create a new division wave in LEC: large teams invest in technology to maintain top positions, smaller teams are left behind. Within the next 2-3 years, we may see an increasingly large gap between the top 4 and bottom 4 of LEC, and this gap will come not only from player quality but also from supporting technology quality. This is a scenario that tournament administrators need to anticipate and prevent.

From a business strategy perspective, GIANTX's decision is fairly sound in the short term. By signing an exclusive contract, they create an immediate competitive advantage and ensure that the 8.5 million euro investment will yield profits through better tournament results. In the long term, however, this decision could backfire if Riot Games decides to tighten regulations or if other teams find ways to build similar tools. This is a classic example of the trade-off between short-term advantage and long-term risk, a lesson any manager must face when investing in new technology.

Contrarian

However, looking at the debate from another angle, there is a strong argument that I believe public discourse is missing. That is: banning or restricting AI coaching tools could cause more harm than good. In football history, whenever new technology emerged, it was initially opposed, but ultimately it became an indispensable part of the game. VAR (Video Assistant Referee) caused intense controversy when first introduced, but now no one can imagine a World Cup without VAR. Goal-line technology went through a similar process. And now, AI coaching may be at the starting point of a similar journey.

This argument has its own strength, but it also has weaknesses that need careful analysis. VAR and goal-line technology are both passive technologies; they provide objective information for humans to decide. AI coaching, on the contrary, is an active technology: it suggests actions, predicts situations, and can potentially completely replace coaches in some aspects of decision-making. This difference is significant. When an AI tool provides tactical suggestions, it is participating in the human decision-making process, not merely providing information. And when decision-making is influenced by AI, the question becomes: where is the boundary between support and replacement? Where is the boundary between humans and machines?

Another perspective worth considering is the role of AI tools in developing smaller teams. While large teams can build their own internal AI tools, smaller teams do not have that capability. If AI tools are completely banned, smaller teams will be on equal technological footing with larger teams, but this also means the entire industry will not benefit from the technology. On the other hand, if AI tools are allowed for widespread use but with clear regulations, smaller teams can access technology through sharing agreements or open-source tools. This is a model that has worked well in many other industries.

The key point of this debate is not whether AI coaching should be used or not, but how to use it fairly. And this is a question that tournament administrators need to answer in the near future, because if they do not answer it, companies like iTero will answer it for them.

The iTero-GIANTX story also raises an interesting question about the future of esports coaching. If an AI tool can provide better tactical suggestions than a human coach, what will be the role of coaches? Some believe AI will completely replace coaches; others believe AI will become a powerful assistant for coaches. Both views have value, but both overlook an important aspect: esports is not only about tactics, it is also about psychology, motivation, and team spirit. These are things that machines cannot replace, at least not yet. And this is precisely why the best coaches will still exist in the future, but average coaches may have to face fierce competition from AI tools.

In the long term, I believe the esports industry will develop in a direction similar to the football industry: AI tools will become an indispensable part of the training process, but there will be clear regulations to ensure fairness. Tools will be classified by the level of intervention in decision-making: from passive video analysis to active tactical suggestions. Low-level tools will be used freely; high-level tools will be restricted or require consent from all parties involved. This is a complex governance model, but it may be the only way to balance technological innovation and competitive fairness.

Takeaway

The 47-minute interview in Reykjavik was not an isolated event. It is a symptom of a larger transition taking place in the esports industry: the transition from competition based on individual skills to competition based on access to technology. In the future, an esports team may lose not because the opponent is better, but because the opponent has better tools. This is a future we need to prepare for, not by banning technology, but by building clear rules for its use. And the biggest question I want to send to readers is: will Riot Games have the courage to act before this technological arms race becomes uncontrollable? Or will we have to wait until a major scandal occurs, when some team is discovered using AI to directly interfere with in-match decision-making, before a set of rules is formed? The answer, based on my experience following matches over the past 11 years, is the latter. And that is the most concerning thing.