iTero, GIANTX and the Commercial Boundary of AI Coaching in Esports

HOOK

Jack Williams did not offer a single performance metric. In the interview titled "Jack Williams on iTero, Giant X, and the future of AI coaching in esports," only two things were stated plainly: iTero is working exclusively with GIANTX, and Williams knows his product will be copied. No sample size, no evaluation methodology, no dataset scale. I read the piece at 23:40, right after closing my LEC summer transfer tracking sheet, and wrote one line in the notes column: an exclusive tool inside a closed league is an advantage that never gets relegated away.

That was the only line in the spreadsheet that night without a number attached.

CONTEXT

Eight years tracking the esports labour market from Busan, and I have watched the coaching department restructure in stages. In 2026, a team had one head coach and a part-time analyst. By 2026, a team at LEC level has a head coach, positional coaches, a data analyst, a performance psychologist, and now a tooling vendor. The vendor is the newest link in the chain, and the one with no governance framework around it.

Patch cadence is a first-order commercial variable here, and it differs completely across titles. Valve ships Dota 2 patches rarely, changes systems at scale, then lets the market stabilise for months. A machine-learning model trained on historical data holds its value across a long window. Riot operates on a biweekly cadence in League of Legends. There, the value of an AI tool shifts from decoding the meta to detecting meta drift faster than opponents, which is a tempo advantage, not a knowledge advantage.

A product that is not tied to any single title would have to tell two contradictory sales stories. If it tells only one, that is the first red flag.

GIANTX, by industry reporting, is an EMEA organisation formed through the merger of Excel Esports and Giants Gaming, competing in the LEC. That is a closed league with fixed member slots and no relegation. In that model, structural advantages do not get competed away across a season the way they do in an open circuit. They compound.

CORE

People ask what I look at before a deal closes. I look at motive, not at price. With a tooling supply contract, motive is even more visible than with a player contract.

The real value of an exclusivity arrangement is not the quality of the model, but the window of information asymmetry it creates. Ten teams in a closed league play the same build. The gap between understanding the meta and understanding it one week earlier does not show up on the standings immediately. It shows up in pick-ban priority, in scrim scheduling, in which team is forced to play first without data. By the end of the season, it shows up as points.

In my own match log, I have recorded pick-ban timestamps and between-game break lengths across more than 200 LEC and LCK series since 2026. That column originally existed to measure how fast coaches make decisions. After this interview, I read it back with a different question: during those ten to fifteen minutes, who has the tool and who does not.

I price this kind of contract by opportunity cost, not by sticker price. Three lines have to be calculated. First, switching cost. If GIANTX's preparation workflow is built around iTero's data model, the cost of changing vendors is not a software fee. It is retraining analytical staff, rebuilding scrim protocols, and losing accumulated institutional memory. That is a lock-in contract, not a software licence. Second, the cost of exclusivity seen from the vendor's side. Selling exclusively means declining revenue from the other nine teams in exchange for a higher price per seat. That trade only makes sense if the premium covers the copying risk Williams himself concedes. Third, resale value. A tool tied to a single customer is valued far below a tool with ten customers, even when first-year revenue is identical.

My spreadsheet is full of formulas, but the answer always sits outside the cell. The cell tells me the price per seat. It does not tell me the termination clause, and that clause is what decides who holds negotiating power in year three.

A credible report needs three signatures: the assistant coach, the agent, and the person in the kitchen. For an analytics tool, those three signatures are the vendor's claim, the team's internal data, and an independent verifier. The interview carries the first signature. The other two do not appear. I am not concluding the product is weak. I am concluding that the performance claim cannot be verified, and in valuation, what cannot be verified defaults to zero.

Patch cadence returns here as the deciding factor. In Dota 2, a tool earns its money through depth of historical modelling. In League of Legends, a tool earns its money through speed of reading change. Those are technically different products. If iTero sells a single product into both ecosystems with one message, that is the question I want asked in the next interview.

CONTRARIAN

The debate the article frames may have picked the wrong centre of gravity. The AI-assisted cheating section reads dramatically, but real-time in-game assistance is already unambiguously banned in every major title. There is no grey zone left to argue about there.

The real grey zone sits in the between-game window of a BO3 or BO5, where the rules have not settled whether a model producing pick-ban recommendations during a break counts as real time. That is ten to fifteen minutes per series, and it is where an AI tool's value peaks, because that is when coaches must decide with the least time available. Current rulebooks were not written for this scenario.

But the larger blind spot is elsewhere. Exclusivity is not a distribution strategy. It is a bet on a governance framework that does not yet exist.

Leagues currently permit exclusive tooling deals because nobody has proven they change competitive outcomes. Once someone does, the operator faces exactly the two options it once faced over in-game coach communication: mandate equal access, or restrict the tool. That means an exclusivity contract carries an implicit expiry date set by the publisher, not by the vendor. Pricing a multi-year exclusivity deal without modelling that risk is the same error clubs made with sponsorship deals that skirted financial fair play.

The pandemic did not kill the transfer market; it merely stripped the camouflage off rules we disguised as FFP. AI tooling will do the same to the coaching market.

TAKEAWAY

I am not arguing that iTero will succeed, and I am not arguing it will fail. I am arguing that the thing worth watching is not model quality. It is the first piece of league rulemaking that governs third-party analytics access. If a publisher writes that clause into its rulebook within the next eighteen months, the entire valuation model for this class of contract has to be rewritten.

And when that happens, the first question I will put to any vendor is this: if exclusivity is stripped away next season, what are you still selling?