For decades, football coaches built their game plans on video sessions, scouting reports and personal instinct. Those tools still matter, but artificial intelligence (AI) is now changing how many staff rooms work. From corner-kick routines to injury warnings and opponent analysis, AI helps coaches process far more information than any human team could handle alone.
This article explains what AI actually does for a coaching staff. It then looks at four key areas: set-piece tactics, open-play analysis, injury prevention and tournament preparation. It ends with the limits that keep human coaches firmly in charge.
What AI Means in a Coaching Context
In football, AI usually refers to machine learning systems that find patterns in large datasets. These include tracking data, which records the position of every player throughout a match, plus event data such as passes and shots, and fitness data from GPS vests worn in training.
A human analyst might spend days reviewing footage of one opponent. An AI model can scan thousands of matches and highlight repeated patterns in minutes. The coach still decides what to do with that information.
Set Pieces: Liverpool and TacticAI
The best-known example of AI tactics comes from Liverpool. Google DeepMind introduced TacticAI in the journal Nature Communications as a system that gives experts tactical insights, especially on corner kicks, after a multi-year research collaboration with Liverpool.
The model was trained on 7,176 corner kicks from Premier League matches, using a technique known as geometric deep learning. In simple terms, it treats players as connected points on a graph. It proved accurate at predicting who would receive a corner and whether a shot would follow, and it could suggest new player positions.
The testing was careful. Five Liverpool experts looked at 50 corner set-ups and chose between an AI suggestion and a real tactic without knowing which was which. They picked TacticAI’s version in 45 cases, or 90 percent of the time. One suggestion involved making sure defenders furthest from the corner made better covering runs.
There was an important caveat, though. At the time of publication, no players had tested the suggestions on the pitch, and an outside expert warned that real value could only be proven in matches.
From Corners to Open Play: Palmeiras
In 2026, TacticAI moved beyond set pieces. On June 11, DeepMind announced that Brazilian club Palmeiras was the first team to build on the system, which can simulate open-play situations up to eight seconds ahead. The deal, revealed at the Google for Brazil event, also involves the Brazilian Football Confederation (CBF).
Open play is much harder to model than a corner, because all 22 players move freely. Palmeiras analysts can drag a player to a new spot on a digital board and watch the model redraw the likely outcome. This lets coaches test “what if” ideas before taking them to training.
Injury Prevention and Workload
AI also shapes squad rotation. Since the 2017-18 season, La Liga side Getafe have worked with California-based AI company Zone7 to analyse performance data and predict when players face injury risk. Clubs including Rangers, Real Salt Lake and Toronto FC have also sent their data to Zone7, receiving daily emails that flag players near a “danger zone.”
The results at Getafe were notable. Head of performance Javier Vidal said the club saw a 40% drop in injury volume in its first year with the system. For coaches, this kind of warning can decide whether a key player starts, rests or trains lighter. However, many accuracy figures in this field come from the companies themselves, so independent evidence remains limited.
Tournament Preparation: FIFA’s Football AI Pro
The 2026 World Cup showed how widespread AI coaching tools have become. FIFA President Gianni Infantino and Lenovo chief Yuanqing Yang revealed Football AI Pro at CES in Las Vegas. It is a generative AI assistant built for all 48 teams in Canada, Mexico and the United States, analysing hundreds of millions of FIFA data points to produce insights in text, video, graphs and 3D visuals.
The goal was fairness. FIFA wanted every team to have the same analysis tools before and after all matches, so smaller federations without large data departments were not left behind. National teams have also used AI independently. England’s Football Association said AI reduced its analysis of opponents’ penalties from five days to about five hours.
Why Coaches Still Make the Decisions
AI is a powerful assistant, but it does not replace coaching. Models depend on good data, cannot fully measure confidence or dressing-room mood, and their ideas still need testing on real grass. A suggestion that looks perfect on a screen may not suit the players available on a cold Tuesday night.
The real shift is in how coaches spend their time. Instead of searching through hours of video, they can focus on teaching, motivating and choosing between better options. As tools like TacticAI and Football AI Pro spread, the advantage may go less to the clubs with the most data and more to the coaches who ask the smartest questions.
