Fantasy Premier League Copilot Companion Brings AI Advice for 2026/27

Fantasy Premier League is adding an AI-powered assistant to one of football’s most demanding digital games, positioning Fantasy Premier League Companion powered by Copilot as a conversational guide for managers navigating player picks, transfers, fixture swings, captaincy calls, and the league’s famously dense ruleset. Built on Microsoft Foundry, using Azure OpenAI, GPT-5.4, and official Premier League data, the new feature is being prepared ahead of the 2026/27 campaign, which begins on Friday, August 21, 2026. Microsoft’s announcement frames it not as an automated team-selection engine, but as a data-backed assistant intended to make FPL’s underlying information easier to discover and interpret.
For the Premier League, this is a consequential expansion of its wider Microsoft partnership. For players, it is an attempt to solve a familiar problem: Fantasy Premier League is easy to start, but hard to master. Millions of managers can choose a squad; relatively few have the time, confidence, or statistical fluency to turn form tables, injury news, team selection patterns, transfers, fixture schedules, and bespoke fantasy metrics into consistently good decisions.
The Companion’s promise is therefore less about finding a magic answer and more about reducing the friction between a question and the official data needed to answer it. That distinction will matter. FPL is built on uncertainty, and a tool that surfaces options, explains trade-offs, and lets people retain ownership of their decisions has far more long-term potential than one that simply tells everyone which player to buy.

Overview: a Copilot for Fantasy Premier League decisions​

The Fantasy Premier League Companion builds on the existing Premier League Companion, an AI-powered conversational experience developed through the league’s partnership with Microsoft. The broader Premier League Companion is intended to operate across the league’s official digital estate, while the FPL-focused layer narrows that capability toward the weekly decisions that define fantasy management. The Premier League’s Microsoft partnership page says the wider Companion draws on more than 30 seasons of statistics, 300,000 articles, and 9,000 videos.
That scale helps explain why a conversational interface may be useful. The modern FPL manager is not short of information. Official pages, fantasy-focused editorial, social media discussion, data visualizations, team news, tactical analysis, and fan-created spreadsheets are available in abundance. The difficult part is knowing what applies to a specific situation right now.
A manager may not want a generic list of high-scoring midfielders. They may instead need to know which midfielders offer the best combination of recent shots in the box, starting security, favorable fixtures, price compatibility, and a realistic route into their current squad. That is the sort of multi-part question the Companion is designed to translate into a readable response, based on the data the Premier League already holds.
According to Microsoft’s feature interview with Premier League senior gaming manager Andy Selby, users will be able to ask broad questions such as who they should consider for their team, or far more specific queries around player categories and performance indicators. The stated aim is to help newcomers understand the game while allowing experienced managers to interrogate data with less manual navigation.
That is a sensible target. FPL is not merely a game of player popularity or season-long points totals. It is an ongoing resource-management challenge, shaped by deadlines, squad structure, opponent strength, injuries, rotations, price changes, captaincy multipliers, and the cost of each additional transfer.

---

Background: why FPL needs a more approachable data layer​

Fantasy Premier League’s basic concept remains highly accessible. Managers select real Premier League players, and those players earn fantasy points according to their performances in real matches. But the apparent simplicity masks a game with numerous interacting systems and a weekly tempo that can quickly overwhelm first-time players.
The Premier League’s official FPL help guide confirms the essential structure:

• Managers have a £100.0 million budget.
• They select a squad of 15 players.
• Only 11 players are chosen to score in a given Gameweek.
• A selected captain receives double points.
• The season is split into 38 Gameweeks.
• Managers can use transfers to react to injuries, form, fixtures, and changing roles.
• Four chips—Wildcard, Free Hit, Bench Boost, and Triple Captain—can radically alter a Gameweek strategy.

Even that list understates the mental overhead. A player who has performed brilliantly in the past may be a poor pickup if their minutes are about to decline. A defender may have attractive fixtures but be vulnerable to losing his starting role. A captaincy decision may look obvious until a manager considers travel, a midweek European match, tactical rotation, or a difficult opponent’s defensive record.
The Companion is positioned as a response to this complexity. Rather than requiring users to locate and interpret separate tables for minutes, tackles, shots, transfers, fixture difficulty, captaincy popularity, and ownership, it is expected to combine that information in a conversational format. Selby’s description of the feature makes clear that its intended audience spans both newcomers asking how to build a team and advanced users looking for answers around narrower statistical criteria.

The importance of the first few Gameweeks​

The early part of an FPL season can be particularly intimidating. Everyone begins with a blank slate, but the decisions made before the first deadline establish a squad’s price structure, depth, captaincy options, and flexibility for early transfers.
The official guide advises managers to construct an initial squad around regular starters, attacking players with goal or assist potential, defenders and goalkeepers with clean-sheet prospects, a playable bench, and one or two reliable captaincy candidates. Those recommendations are logical, but they still leave room for dozens of judgment calls.
This is where an official FPL AI assistant could be most valuable. A newcomer might ask why a cheaper player who starts every week can be more useful than a higher-profile substitute. They may want an explanation of why a premium forward makes the rest of the squad harder to build. They may be confused by why ownership percentages are interesting but should not be treated as proof that a player is the correct pick.
The Companion could serve as a practical bridge between the rules and strategy. It could explain the why behind the advice, not just repeat a template squad or a list of last season’s stars.

A season with a later start​

The timing is also notable. The 2026/27 Premier League season starts on August 21, 2026, one week later than the usual mid-August opening. The league says the schedule provides 89 clear days after the prior season and 33 days after the FIFA World Cup 26 final, while the campaign itself will include 33 weekends and five midweek match rounds. The Premier League’s season-date announcement puts player welfare and calendar congestion at the center of that decision.
For FPL managers, this context has practical implications. A summer tournament, transfers, changing tactical systems, and delayed player returns can make opening squads more volatile than usual. A tool that can help distinguish official information from stale assumptions may prove more useful than one that simply produces pre-season projections.

What the Companion is expected to do​

The clearest definition of the Fantasy Premier League Companion is that it is an advisory interface, not an autopilot. The Premier League has deliberately drawn a boundary between helping managers make choices and making those choices on their behalf.
Microsoft’s report says the Companion will not predict FPL points or run a user’s team. Instead, it will offer multiple options, provide relevant supporting data, and connect managers to related editorial where appropriate.
That policy is a strength, not a limitation.

Advice without surrendering agency​

FPL is rewarding precisely because decisions carry consequences. A successful differential transfer, a disciplined decision to avoid a popular player, or a captaincy call that beats a mini-league rival feels meaningful because the manager made it.
An automatic recommendation engine can create the opposite dynamic. If every user receives one supposedly optimal answer, it risks turning the game into a race to accept the same suggestion first. The best fantasy games preserve room for competing interpretations of the same evidence.
The Companion’s planned approach appears better aligned with that reality. If a manager needs to replace an injured player, the system can surface several alternatives and explain the statistical or schedule-based case for each. One option may be more secure for starts; another may have stronger short-term fixtures; another may represent a different budget tier. The manager still decides which trade-off matters most.
This is especially important because football cannot be reduced to a predictable spreadsheet. A strong fixture can still end in a blank. A player with excellent underlying statistics can be substituted early. A team can change system, suffer an unexpected injury, or prioritize another competition. The Companion’s refusal to present itself as a fortune-telling service is an appropriate acknowledgment of those limitations.

Using both match data and game data​

The most interesting technical and product decision is the plan to combine two distinct categories of information.
First, there is match data: goals, assists, appearances, cards, minutes, tackles, shots, and other indicators from Premier League fixtures. Second, there is FPL game data: transfers, captaincy selections, chip usage, player popularity, and other signals that exist because millions of managers are playing the fantasy game. Selby’s explanation describes the Companion as bringing those two sets together for recommendations.
That combination makes the feature more useful than a basic football chatbot. Match data can identify performance and involvement. FPL data can reveal market behavior, price pressure, transfer trends, and the extent to which a player has become a popular captaincy or ownership choice.
Neither type of data should be used in isolation:

• Match data alone can miss the practical FPL context, including price, ownership, and transfer momentum.
• FPL data alone can become an echo chamber, with managers chasing each other into short-lived trends.
• The combined view can expose the difference between a popular choice and a well-supported one.

For example, a manager asking whether to sell an underperforming midfielder may need more than a recent-points comparison. The better answer could incorporate minutes played, attacking volume, opponent quality, upcoming fixtures, transfer activity, and feasible replacements within the existing budget.

Real-time relevance, with an important caveat​

The league says that new information can alter the Companion’s answers during the week, meaning a query on Tuesday could produce a different response by Thursday if team news or other relevant updates arrive. Microsoft’s report presents this as a key advantage of the conversational system.
That is exactly what managers need ahead of Gameweek deadlines. FPL decisions are perishable. An answer that was sensible before a press conference, a training-ground update, or a transfer announcement may be less useful afterwards.
However, real-time should not be mistaken for infallible. The speed of new information makes provenance essential. Managers should still distinguish between official club confirmation, reliable reporting, opinion, tactical speculation, and a player’s own social-media activity. The Companion’s value will rise significantly if responses make clear which facts are confirmed and which conclusions are interpretation.

---

The Microsoft Foundry foundation​

For Windows and enterprise technology readers, the football story is also an instructive example of how a major consumer-facing AI product can be assembled from enterprise cloud components.
The Companion is built using Microsoft Foundry, the company’s platform for developing, deploying, managing, evaluating, and governing AI applications and agents. Microsoft’s Foundry documentation describes it as a unified environment for agents, models, tools, observability, identity controls, networking, and policy management.
This does not mean every FPL recommendation is simply a model-generated opinion. The intended architecture matters: language models handle the conversational layer, while official Premier League and FPL information serves as the knowledge base that gives answers their practical grounding.

Why grounded data is the central issue​

A language model is capable of writing fluent and persuasive answers even when its information is incomplete, outdated, or wrong. Sports data creates a particularly difficult version of this problem because facts can change quickly: team news changes, roles change, injuries occur, fixture schedules shift, and player transfers rewrite the assumptions behind an answer.
The appropriate solution is not to treat the model as an authority in isolation. It is to use it as an interface over controlled sources, structured statistics, and timely editorial material.
The Premier League’s stated use of official data is therefore the most important detail in the announcement. It provides a path toward answers that do more than sound convincing. If implemented carefully, the Companion can direct managers toward meaningful evidence—such as minutes, shot volume, transfer trends, or fixture context—rather than asking them to trust an unexplained recommendation.
Microsoft Foundry is designed to support exactly these kinds of production requirements. Its documentation highlights built-in capabilities for tracing, monitoring, evaluation, role-based access control, policy controls, content filtering, and network isolation. Those platform features are not visible to the average FPL manager, but they are relevant for a service that may face enormous demand around weekly deadlines.

GPT-5.4 and conversational reasoning​

Microsoft says the FPL Companion uses GPT-5.4, a model that OpenAI has made available through both ChatGPT and its API. OpenAI’s GPT-5.4 announcement describes it as a mainline reasoning model with tool-use capabilities and API availability under the gpt-5.4 model name.
In the FPL setting, the value of such a model is not that it can predict the future. Its practical advantage is language understanding. Managers rarely frame questions in clean database terms. They ask things like:

• “I need a midfielder under a certain price who is likely to start.”
• “Should I save my transfer or replace this injured defender?”
• “Which captain has the better combination of form and fixture?”
• “Explain whether a points hit is worth it in this situation.”
• “Who are the alternatives if I do not want to follow the most popular transfer?”

A capable model can interpret the intent behind those questions, retrieve the appropriate data, and return an answer in plain English. It can also maintain the context of a conversation, so a follow-up request does not need to repeat every detail.
Microsoft Foundry’s current model catalog includes the GPT-5.4 family among Azure OpenAI offerings, alongside other model lines and deployment choices. Microsoft’s model documentation also underscores a broader industry point: the model is only one component. Deployment, data processing, authentication, filters, evaluation, and operational governance all shape whether an AI feature is dependable in the real world.

---

The strongest use cases for FPL managers​

The Companion will need to prove that it adds value beyond a traditional search box or a static player-ranking page. The best early use cases are likely to be practical, repeatable, and tied directly to familiar FPL pain points.

1. Explaining FPL rules in plain language​

The official FPL rules are detailed, and even experienced players can be caught out by mechanisms such as auto-substitutions, free-transfer rollovers, chip availability, captaincy delegation, and the deadline system.
The current official guide states that managers receive one free transfer per Gameweek, can roll unused transfers up to a maximum of five, and lose four points for additional transfers made in the same Gameweek. The rules explainer also confirms that managers have two sets of chips—one for each half of the season—and unused chips do not carry forward.
A Companion that can translate those rules into scenario-specific guidance could remove a large barrier for new users. “Can I use this chip now and save the other for later?” is more useful when answered against the manager’s actual situation than when buried in a long FAQ.

2. Identifying replacements after injuries or suspensions​

A player injury is one of the most common triggers for an FPL transfer, but selecting a replacement requires several filters at once. Is the replacement fit? Is he likely to start? Does his team have favorable fixtures? Does he fit the budget? Is the immediate transfer worth making, or is it better to wait?
The Companion’s ability to combine player performance and FPL game data could make this process substantially quicker. It should be able to offer a list of candidate replacements, explain the differences, and point out any assumptions rather than pretending that one answer is universally correct.

3. Making fixture-based selections easier to understand​

Fixtures have always been central to FPL strategy, but their effect can be misread. A supposedly easy opponent may be defensively strong away from home. A difficult fixture may still suit an elite attacker with penalty duties or set-piece involvement. A team with a short-term favorable run may have heavy rotation risk.
A high-quality conversational answer can introduce that nuance without forcing the manager to manually assemble five separate tables. The key is to explain why a fixture run is appealing and where the counterarguments lie.

4. Helping managers use their chips with more discipline​

Chips can transform an FPL season, but they are frequently wasted through impulse decisions. Triple Captain, Bench Boost, Free Hit, and Wildcard all involve opportunity cost; using one in an ordinary Gameweek means it cannot be used when the schedule becomes more favorable.
The Premier League’s official help page confirms that only one chip can be played in a Gameweek and explains the different effects of each chip. The same guide shows why an assistant could be useful: the rules are straightforward in isolation, but their strategic consequences are not.
A sensible Companion response should avoid overconfident commands. Instead of saying “use your Wildcard now,” it should frame the decision around squad weaknesses, expected future schedule disruptions, injuries, fixture runs, and how many transfers would otherwise be needed.

5. Supporting healthier mini-league competition​

Fantasy football is a social experience as much as an analytical one. The official FPL guide emphasizes that players can create and join leagues with friends, family, and colleagues, alongside broader public competitions. That social layer is often what keeps casual managers engaged after a bad Gameweek.
The Companion could help by making FPL less exclusionary. A new participant in an office league should not need years of experience with fantasy jargon to understand a captain pick or a transfer. Better onboarding may create closer, more enjoyable competition rather than merely helping expert managers find marginal edges.

---

Risks: the limits of AI-assisted fantasy management​

The potential is clear, but the Companion’s success will depend on how carefully it avoids several predictable problems.

Overconfidence is the biggest product risk​

AI answers can sound definitive even when the underlying decision is uncertain. In FPL, that is dangerous because the game is full of variance. A player can have a poor performance after excellent preparation, and an apparently irrational choice can succeed because football is not deterministic.
The Premier League has already made the right philosophical choice by saying the Companion will not predict points or act as autopilot. Its stated position protects the fundamental manager-versus-manager nature of FPL.
But the user experience must reinforce that principle. Responses should use calibrated language such as “consider,” “based on recent data,” “a short-term option,” and “a risk to monitor.” A tool that presents probability as certainty will damage user trust, especially after inevitable bad outcomes.

Data freshness must be transparent​

If a response changes because new information has arrived, that is useful. Yet it also creates an obligation to show when the data was updated and which information drove the change.
A deadline-day recommendation is not the same as a midweek recommendation. A manager needs to know whether the answer incorporates the latest confirmed injury news, a recent transfer, or a club announcement. Without that clarity, “real-time” can become an opaque marketing term rather than a meaningful advantage.

Popularity can become self-reinforcing​

FPL game data such as transfers and captaincy selections can be valuable signals, but they can also amplify herd behavior. A player’s rapid rise in popularity may reflect sound analysis—or it may reflect anxiety, social-media momentum, and fear of missing out.
The Companion should therefore treat popularity as context, not a verdict. An especially useful response would explain:

• Whether a player is attracting transfers.
• What recent football data supports that movement.
• What risks remain despite the trend.
• Which lower-owned alternatives have comparable credentials.
• Whether the manager’s own squad structure makes the popular move sensible.

This would make the assistant more than a sophisticated bandwagon tracker.

The platform must remain fair​

There is also a broader competitive issue. Any official tool that delivers insight must avoid creating a divide between people who can formulate elaborate prompts and those who cannot. The solution is not to remove advanced functionality; it is to ensure that the interface works equally well for simple questions and that core information remains accessible outside the chatbot.
The best version of the Companion will help someone ask, “Who should I captain?” without shaming them for being new, while also letting an expert investigate a much more detailed question. That layered design is consistent with the Premier League’s stated goal of helping players at different experience levels. Selby’s comments explicitly place new and advanced managers within the intended audience.

---

What this means for the future of sports AI​

The Fantasy Premier League Companion is part of a larger shift in sports technology. The first generation of digital sports products brought fans more content. The second brought them more data. The emerging generation is trying to make data interactive, allowing a supporter to ask a question in their own terms and receive a tailored explanation rather than a static dashboard.
That can be genuinely useful when the assistant is grounded in official information, transparent about uncertainty, and designed to complement—not replace—human judgment. FPL is an especially suitable proving ground because its audience already thinks in terms of evidence, risk, player availability, and short-term decisions.
Microsoft and the Premier League have an opportunity to demonstrate a stronger model for consumer AI than the familiar “ask anything” chatbot. The better proposition is narrower and more disciplined: ask FPL-specific questions, receive evidence-based context, and make the final call yourself.
For the millions who assemble a £100 million virtual squad each season, the Companion may not guarantee a green arrow, rescue a failed captaincy call, or reveal the perfect differential. It could, however, make the game’s considerable depth more accessible, reduce the time required to find trusted information, and help managers understand the reasoning behind their choices. That is a much more credible definition of AI-powered fantasy football success.

---

References​

1. Primary source: Microsoft Source
   Published: 2026-07-27T11:10:09.345554
   
   

   Fantasy Premier League Companion gives managers a new tool for success - Source EMEA

   news.microsoft.com
2. Related coverage: premierleague.com
   
   

   What are the dates for 2026/27 Premier League season?

   All you need to know about when the new Premier League campaign will start and finish

   www.premierleague.com