Technology

Marquee vs SICS Atlas:What a Football Chat Layer Can and Cannot Do

12 min 12.09.2026

A new product category has appeared in football analytics: the chat-native decision layer. Instead of a dashboard of vendor metrics, you type a football question and get an answer. Two names now define that category — Marquee (themarquee.ai) and SICS Atlas from Soccerment — and they are not the same product wearing different logos.

This article takes both products at their word: everything below comes from the vendors' own product descriptions. The interesting question is not whether the claims are true. It is what job each product is built for — and which jobs remain the club's own once the chat window closes.

The Marquee and SICS Atlas logos side by side

Quick comparison

What it is — Marquee: Recruitment decision layer: twelve proprietary AI/ML models + Marq chat · SICS Atlas (Soccerment): Proprietary scouting + video intelligence platform · GAMECODE.Ai: Club operating system: ingest, club-authored metric trees, video workbench, private assistant

Primary job — Marquee: Help a recruitment team decide · SICS Atlas (Soccerment): Watch, telestrate and rank from Soccerment’s own data · GAMECODE.Ai: Compile the club’s game from any feed and any camera

Data foundation — Marquee: 90+ licensed competitions — event, physical, off-the-ball, financial — plus the club’s own stack, data-agnostic · SICS Atlas (Soccerment): 310+ competitions proprietary; 6,500+ matches with video per month; 50 competitions with tracking; XSEED GPS · GAMECODE.Ai: Whatever the club owns or buys: event, tracking, cameras, chips — practices and historic tape included

Club fit / game model — Marquee: Sporting-DNA profile scored by xFit (System Fit + Squad Fit; Projected Output coming soon) · SICS Atlas (Soccerment): Soccerment’s metric catalogue: 90k players × 100+ metrics · GAMECODE.Ai: The club authors the algorithms themselves — Studio metric trees the staff own

Video — Marquee: Recruitment reports and chat; video is not the centre of the product story · SICS Atlas (Soccerment): Real clip workflow: telestration, charts that jump to a clip, on the Atlas library · GAMECODE.Ai: Workbench / holodeck / overlays of the club’s own metrics on any tape

Connected providers named — Marquee: SkillCorner (14 Jul 2026) + Gradient Sports (9 Jul 2026), plus the club’s existing stack · SICS Atlas (Soccerment): Proprietary Soccerment data; XSEED GPS as its own SKU · GAMECODE.Ai: Event, tracking, cameras and chips — DFL/Sportec, TRACAB, Catapult, STATSports, SkillCorner, StatsBomb, Wyscout, KINEXON, Impect, Genius, Footovision, PlayerData, Transfermarkt and more

Chip tracking / streaming — Marquee: Recruitment-data focus · SICS Atlas (Soccerment): XSEED GPS as a separate product · GAMECODE.Ai: Chip tracking (KINEXON etc.) up to 100 Hz; live match streaming is a first-class ingest path

Audit trail — Marquee: Chat and report history · SICS Atlas (Soccerment): Platform records of clips and charts · GAMECODE.Ai: Every ingest and manipulation cryptographically sealed — tamper-proof, fully explainable end to end

Youth / medical / consent — Marquee: Recruitment scope · SICS Atlas (Soccerment): Scouting scope · GAMECODE.Ai: Shared only at will, with full privacy — built for academies holding under-16 talent

Pricing — Marquee: On request · SICS Atlas (Soccerment): On request · GAMECODE.Ai: On request (waiting list)

_Product descriptions from each vendor’s own pages, September 2026._

Marquee

Marquee is a recruitment decision layer. The onboarding captures the club’s sporting DNA — style of play, strategic objectives, financial reality, the archetype wanted in every position — and recalibrates it within hours when the football changes. Twelve proprietary AI and ML models read event, physical, off-the-ball and financial data across 90+ licensed competitions, alongside whatever the club connects from its own stack, and score everything against that DNA.

The output layer is genuinely thought through: Marq answers football-language questions with ranked shortlists; Player 360 builds one report per player across performance, style, squad context and deal context; Opportunity Radar watches contracts, minutes and club situations and filters what is realistically signable; Club Hub turns the same lens on your own squad; Opposition Hub reads the next opponent. Every player carries an xFit score — System Fit and Squad Fit today, Projected Output coming soon — and each component stays visible under the number.

Strengths

A serious recruitment framing. The product sits where a sporting director already lives: in a conversation about who to sign, with the club’s own context in the sum. Two clubs running the same search get different lists — that is the right ambition.

League-difficulty honesty. Players are graded against their own league first, converted on a scale learned from players who actually moved. A standout in a smaller market is not discounted by reputation.

Low friction. It works with the club’s existing stack or with nothing connected — a real buying argument for a club that does not want an integration project before the first answer.

Transparent reports. The three xFit questions stay visible under the score, and the reports say so where the evidence is thin.

Where its job ends

None of this is a criticism — it is the product’s own scope, and scope is what a buyer should understand:

The game model is a profile, not an algorithm the club authors. The sporting DNA configures Marquee’s twelve models. That is powerful for recruitment, and it is a different thing from a Studio in which your staff write and own the metric trees that encode your football.

It reads data that already exists. Licensed packs plus the club’s connected sources. Creating new data — turning Tuesday’s training film or a 2019 cup tie from a hard drive into the same language as Saturday — is a different category of product.

Video is not the product’s centre. The reports are written in scouting language; overlaying the club’s own metrics on the club’s own clips is a different discipline.

Its history is a decision record, not a data-provenance record. Chat and reports document what was asked and recommended. Replaying every computational step from raw feed to final number, for an auditor or a guardian, is a job the product does not set out to do.

Ideal users

Recruitment teams that want a conversational layer with the club’s context built in, over licensed data and their existing stack — and who keep capture, club-authored algorithms and video overlays as separate concerns.

SICS Atlas (Soccerment)

SICS Atlas is not a Wyscout wrapper. Soccerment runs its own stack: 310+ competitions, 6,500+ matches with video per month, 90k players × 100+ metrics, and 50 competitions with tracking. XSEED, Soccerment’s GPS wearable, is its own product line.

Atlas pairs that library with a real clip workflow: telestration, and charts that jump straight to the video behind them. That is genuine in-product video on Soccerment’s matches.

Strengths

A real scouting library with video. 6,500+ matches with video a month is a working product at serious scale.

Tracking on 50 competitions is a substantial positional layer, owned by Soccerment end to end.

Telestration and chart-to-clip is genuine analyst workflow — analysts who live in clips will feel the difference immediately.

Breadth of proprietary data that a single-club compiler does not try to replace.

Where its job ends

Again — scope, not criticism:

The metrics are Soccerment’s. A catalogue of 100+ metrics across 90k players is a scouting instrument. Encoding your game model in algorithms your staff author and own is a different job.

The video is the Atlas library. Matches Soccerment captures and indexes. Practice sessions and the historic tape on your own drives belong to a different workflow.

XSEED is its own product. GPS wearables sold as a SKU, serving its own use case — distinct from streaming chip and match feeds into the same platform that computes the club’s metrics.

Its records document analyst work — clips made, charts drawn. A cryptographically sealed trail from ingest through every manipulation to the number is a different discipline.

Ideal users

Recruitment and analysis teams who want Soccerment’s data and a real clip/telestration workflow at catalogue scale, with the vendor’s metric definitions as the shared language.

Where GAMECODE.Ai fits

Both products above are decision layers on top of data that already exists. GAMECODE is the layer underneath: it compiles the club’s game — and then lets you talk to it.

In priority order:

Ingest any video and interpret it — practices _and_ historic recordings, not only prefabricated event packs. The competitors read data that exists; GAMECODE’s first job is making data exist, from your cameras, your tape, your chips.

Fully explainable. GAMECODE clients put algorithms on the platform (Studio metric trees). A DNA profile tunes a vendor’s models; a metric tree is the club’s model. When the number is later questioned, you open the tree, not a conversation log.

Deep video visualisation of the club’s own metrics. Workbench, holodeck, overlays of _your_ rest-defence metric on Tuesday’s training tape. Atlas’s clip workflow serves Soccerment’s library; this serves yours.

The rest of the football job — event, tracking, cameras, opposition, training and match in one language — already documented on the other comparison pages.

A long connected-provider list: DFL/Sportec, TRACAB, Catapult, STATSports, SkillCorner, StatsBomb, Wyscout, KINEXON, Impect, Genius, Footovision, PlayerData, Transfermarkt, and others already on the public about page.

Chip tracking up to 100 Hz, and streaming. KINEXON-class chips at 20 and 100 Hz are a first-class ingest, and live match streaming (DFL-style feeds) is a platform path. To be precise about our own claim: the platform is built to stream, and 100 Hz chip files already land on the same sealed, auditable record — not that every customer runs a live chip feed today.

And the condition that makes any answer worth acting on:

Always fully explainable. A GAMECODE answer carries its own history: which feeds went in, which of the club's algorithms ran, and every manipulation on the way to the number — cryptographically linked, similar to a blockchain, without mining or tokens. Medical, sleep, nutrition and academy performance records for under-16 talent keep full privacy and are shared only at will. A recruitment recommendation and a telestrated clip answer different questions — both useful, and neither sets out to be the audit trail behind the number. That trail is what GAMECODE is.

What GAMECODE does not do

We do not replace Soccerment’s 310-competition library, and Marquee’s recruitment chat is a serious product for the job it was built for. If you need a huge pre-indexed scouting catalogue, Atlas is built for that. If you need a conversational layer over licensed data with the club’s context built in, Marquee is built for that. GAMECODE is the compiler and the system of record for the club that already has tapes, feeds, chips and a game model it intends to own.

Field notes: chat is the interface, not the product

We talk to our own cube in football language too — that is GC:Pulse. The difference is what the sentence is standing on. Pulse answers a question against your tree, your video, your sealed ingest. Marquee’s story is a recruitment chat with the club’s context over licensed competitions. Atlas’s story is Soccerment’s catalogue plus a real clip tool. Both are legitimate products. Ours is the layer underneath them all: the club’s own ingest, the club’s own algorithms, the club’s own video — and a record you can replay.

Hudl remains a cooperative relationship on the event-data side (see the Opta / StatsBomb / Wyscout article). Marquee and Atlas sit in a different corner of that landscape: the conversational layer above the data.

Use cases

The right tool for each job

Best for a recruitment chat with the club’s context built in, over licensed data: Marquee.

Best for a proprietary scouting library with real telestration: SICS Atlas.

Best for compiling the club’s own game from any tape, any feed and any chip, with algorithms the staff authored and a record you can audit: GAMECODE.Ai.

Best for under-16 / medical / sleep / nutrition on the same platform as the football numbers, shared only at will: GAMECODE.Ai. That guarantee is the platform’s core, not an add-on — and it is a different job than either decision layer sets out to do.

Which should you choose?

If you want a conversational recruitment layer scored against your club’s DNA, on licensed data you may not even connect yet: Marquee.

If you want Soccerment’s library and a clip/telestrate workflow: SICS Atlas. It is a real video product at catalogue scale, with the vendor’s metric definitions as the shared language.

If the question is “can this answer survive the sporting director, the doctor, the guardian and the auditor?” — that is not a chat-UX problem. It is a chain-of-custody problem, and it is exactly what GAMECODE.Ai is for: ingest any video, put _your_ algorithms on the platform, overlay _your_ metrics on the clip, stream and chip-track at match frequency, and keep every step fully explainable.

While other clubs ask a chatbot for a ranking, GAMECODE.Ai gives you a number you can replay — join the waiting list to see it on your own feeds.

The Opta by Stats Perform, hudl|statsbomb and Wyscout logos side by side

Opta vs. StatsBomb vs. Wyscout:Which Football Data Provider Is Best?

10 min 24.07.2026

The Veo, Pixellot and Spiideo logos side by side

Veo vs. Pixellot vs. Spiideo:Which Football Camera System Is Right for Your Club?

11 min 23.07.2026

Show all articles