WOS

AI System - AI Event Feed (Live Stories for Sports Events)

J
João NevesinWOS
7 months ago

Goal
Create a new top-of-feed experience where users can follow live sports events (like matches, races) and get a real-time, AI-curated feed tailored to that event and to the user.


Core concept

At the top of the feed:

  • Show event bubbles (stories-like UI)

    • Circles or squares (like Instagram stories)

    • Each represents a live or upcoming event

Users can:

  • Tap an event → open Event Feed

  • Create a new event (if not listed)

  • Follow/unfollow events


Event Feed (main experience)

Each event has a dedicated feed:

Content types inside

  • Internal articles

  • Social posts

  • Videos

  • Live match data (JSON provider)

  • Commentary / match events

  • External signals (X / selected users)


AI Agent system (core engine)

What the AI does

Agents continuously:

  • collect content from all sources

  • score relevance

  • rank and filter noise

  • update the feed in real-time


Inputs

  • Match events (goals, cards, incidents)

  • Latest news

  • Internal content

  • Social signals

  • User behavior (preferences, teams followed, language)


Output

A ranked stream of content, combining:

  • real-time updates

  • contextual news

  • highlights

  • videos

  • social reactions


Personalization layer

Each user sees:

  • the same event base feed

  • but ranked differently

Based on:

  • teams followed

  • content consumption history

  • preferred language

  • creators followed

  • engagement patterns


Event creation logic

Sources of events

  • Pre-loaded from sports provider (fixtures)

  • User-created events:

    • “Liverpool vs Arsenal”

    • “F1 Monaco GP”

When created:

  • event becomes globally available

  • other users can follow it


Event structure

Each event has:

  • event_id

  • sport

  • teams / participants

  • start_time

  • status (upcoming / live / finished)

  • tags (teams, competition, country)


UI structure

1. Top feed (Stories-like)

  • horizontal scroll

  • live indicator (red dot / pulse)

  • upcoming vs live distinction


2. Event page tabs

Feed

  • AI curated content stream

Live

  • structured match data (timeline, stats)

Discussion

  • user comments / chat (threaded or real-time)


Discussion layer

Users can:

  • comment on events

  • reply to others

  • react

Creates a live conversation layer around the event.


Ranking logic (simplified)

Each content piece gets a score:

score = relevance + freshness + event_signal + engagement + personalization

Where:

  • relevance → match/event related

  • event_signal → triggered by match events (goal = boost)

  • engagement → likes, comments, views

  • personalization → user affinity


Real-time behavior

  • feed updates dynamically during live events

  • key moments trigger boosts:

    • goals

    • red cards

    • race incidents

  • AI prioritizes high-signal moments


Data layer

  • central event table

  • event_content mapping table

  • real-time ingestion pipeline

  • user-event preferences


Rollout plan

Phase 1

  • Event bubbles UI

  • Static event feed (non real-time)

  • Basic AI ranking

Phase 2

  • Real-time updates

  • Live match data integration

  • Personalization layer

Phase 3

  • Discussion/chat

  • External sources (X integration)

  • Advanced AI scoring


Outcome

This becomes one of the most powerful features in WOS:

  • turns the platform into a live sports companion

  • increases session time dramatically

  • creates real-time engagement loops

  • connects content, data, and community in one place

This is not just a feature — it’s a new core product layer for WOS.

  1. In Review
  2. Planned
  3. In Progress
  4. Completed
Loading content, please wait...
Loading content, please wait...
Loading content, please wait...
Loading content, please wait...
Loading content, please wait...
Loading content, please wait...
Loading content, please wait...
Loading content, please wait...
Loading content, please wait...
Loading content, please wait...
Loading content, please wait...
Loading content, please wait...