Maya Social Fabric & Community Health Proposal

Building MMO Community Systems That Help Players Find Their People#

Date: 2026-03-19 Scope: Maya multiplayer platform, Lilith companion surfaces, Sophia analytics, Hathor world design, Maya engine runtime Status: Proposal

Table of Contents#

Executive Summary
Industry Research Synthesis
Design Principles
System Proposal
Architecture Recommendations
Implementation Priorities
Success Metrics
Risks and Anti-Patterns
Reference Links

1. Executive Summary#

Phase 77 gives Maya the technical skeleton of a world-class multiplayer platform: transport, matchmaking, lobbies, parties, guilds, communication, cross-platform identity, orchestration, live events, and networked AI.

What it does not fully cover is the difference between:

a multiplayer system that lets people connect
and a social ecosystem that reliably helps people belong

The strongest MMO and live-service communities do not grow from guild chat and party lists alone. They grow when the game repeatedly helps players:

meet compatible people
move from solo to duo to group to guild with low friction
contribute before committing
find welcoming sub-communities quickly
recover from bad experiences
return after time away without feeling socially reset
receive human and AI guidance that lowers the social cost of participation

Deep review of current industry patterns shows the most effective community platforms combine:

structured search and discovery
low-pressure group formation
persistent identity spaces
shared goals and contribution systems
newcomer and returner lanes
proactive trust and safety
strong out-of-game companion surfaces
social-health instrumentation
increasingly, AI or agentic assistants that help players orient themselves and take the first step

This proposal introduces a dedicated community phase focused on social formation, guild growth, belonging, moderation, and AI-assisted social facilitation. It is deliberately not just a restatement of "friends, parties, guilds." The gap is deeper: the game needs systems that make social success more likely, especially for shy, awkward, isolated, or returning players.

The design target is an MMO/multiplayer ecosystem where:

a shy player can find a comfortable first group
a solo player can be gently routed toward recurring teammates
a guild can grow through shared work, not just spam recruitment
a returner can quickly re-enter the social graph
AI companions can act like transparent, opt-in social scaffolding rather than fake human substitutes
NPCs and world spaces can help coordinate, host, and lower social friction

2. Industry Research Synthesis#

2.1 What Leading Games and Platforms Repeatedly Do#

The same ideas appear across major MMOs, live-service games, and community platforms:

Structured listings beat vague chat spam. Bungie's Fireteam Finder uses searchable listings with tags, scheduled start times, application vs auto-join settings, platform preferences, language expectations, minimum rank, and a ready-up flow.
Cross-world recruitment meaningfully expands the social pool. FFXIV's Duty Finder, Party Finder, cross-world private parties, and join-in-progress support reduce the likelihood that a player gets socially stranded on one shard or server.
Searchable recruitment pages matter. WoW's Guild & Communities Finder and FFXIV's Free Company search both expose filters for focus, roles, activity, size, and recruiting status.
Players want multiple layers of belonging. WoW Communities, FFXIV linkshells/free companies, Destiny clans, EVE corporations, and Warframe clans/alliances all support different social depths rather than forcing a single "one true guild" model.
Shared work is better than generic chat. EVE Corporation Projects and Freelance Jobs show that task boards, contribution tracking, rewards, deadlines, visibility rules, and newcomer-friendly low-commitment work are powerful social glue.
Shared spaces create identity. FFXIV estates, Warframe dojos, and similar guild spaces act as communal memory, status display, coordination hubs, and onboarding surfaces.
Healthy communities are explicitly designed, not merely moderated. Riot's Player Dynamics work argues that better social outcomes come from changing the environment and reducing friction, not only punishing bad actors after the fact.
Onboarding must be personalized. Discord's community onboarding, rules screening, welcome flows, and channel/role personalization show that large communities remain usable only when new members are not dumped into every surface at once.
Safety must scale proactively. Discord AutoMod, alert channels, and server insights demonstrate the value of pre-post filtering, anomaly detection, escalation channels, and health analytics.
Social graphs need to stay current in real time. Xbox Social Manager and PlayFab community services show the importance of fresh presence data, filtered groups, durable player groups, searchable lobbies, and cross-network friend visibility.
AI-guided onboarding is now practical. EVE's AIR Career Program uses Aura to suggest activities, skill plans, and freelance jobs; NVIDIA ACE examples show companions that can find loot, suggest tactics, accompany the player, and act like co-playable teammates.
LLM agents can coordinate believable social activity. The Generative Agents research demonstrates memory, planning, and reflection systems that can generate emergent invitations, introductions, and event coordination.

2.2 The Core Industry Techniques and Tricks#

Below are the patterns that consistently differentiate sticky communities from social ghost towns.

2.2.1 Lower Commitment Before Asking for Commitment#

Strong communities let players progress through smaller commitments:

watch activity
browse a listing
join a one-off run
join again
become a provisional member
become a trusted member
become a host, mentor, or officer

Asking a new player to join a permanent guild before they have even enjoyed one good run is usually premature.

2.2.2 Match on Intent, Not Just Skill#

Skill/MMR is only one axis. The strongest social matches also consider:

role needed
schedule overlap
language
voice comfort
activity intensity
teaching willingness
first-timer friendliness
reliability
play session length
personality/energy compatibility

A "good teammate" for community growth is often not simply the highest-skill teammate.

2.2.3 Build Repeated Contact Loops#

Friendship requires repeated contact. Systems should increase the chance that compatible players see each other again through:

follow-up prompts after a successful run
"play again tomorrow" scheduling
automatic re-queue together
recurring guild/project groups
shared event calendars
return-to-previous-squad suggestions

2.2.4 Give Players Meaningful Ways to Contribute Before They Belong#

EVE's project and freelance structures highlight a powerful idea: players can start helping before they fully join.

That implies MMO community systems should support:

guest contributions
alliance guest passes
provisional memberships
open guild projects
public help boards
newbie tasks that visibly matter

Communities deepen when the world has places designed for meeting:

starter taverns
campfires
guild open houses
training grounds
crafting plazas
event stages
fishing docks
bulletin boards

Menus recruit. Places create culture.

2.2.6 Treat Moderation as Infrastructure#

Healthy communities are easier to build when:

newcomers accept clear rules before speaking
admins can assign channels based on interest
automation blocks obvious abuse before it lands
moderators have clear escalation tooling
health metrics reveal degradation early

2.2.7 AI Should Scaffold Human Community, Not Replace It#

The most promising AI role is not "fake friend."

It is:

guide
translator
coach
concierge
summarizer
connector
social icebreaker
event host assistant
missing-slot companion

AI is strongest when it lowers the cost of human connection and hands the player off to real people.

3. Design Principles#

3.1 Human-First, AI-Assisted#

The goal is not to keep players attached to AI indefinitely. The goal is to use AI and world systems to help players become comfortable enough to engage with other humans.

Players need clear social pathways:

solo
solo plus AI companion
duo
temporary group
recurring group
guild/community
alliance/network

The system should never force a player to jump from isolated solo play directly to high-pressure guild participation.

3.3 Shy-Player Safe by Design#

The platform should include low-pressure paths for players who are:

socially anxious
inexperienced
new to MMOs
new to voice chat
new to a game mode
returning after a long absence

That means quiet lanes, first-timer tags, optional mic, teaching-friendly groups, AI rehearsal tools, and world spaces where simply being present feels acceptable.

AI systems must:

disclose that they are AI
operate only with clear opt-in for socially meaningful interventions
respect privacy, blocks, and invisibility settings
avoid deceptive or manipulative emotional design
never impersonate a human player

3.5 Reward Prosocial Behavior Without Turning It into Grind#

Recognition matters, but badly designed rewards create performative or abusive behavior. Incentive systems should celebrate hosting, mentoring, welcoming, and reliability without becoming easy to farm.

3.6 Measure Belonging, Not Just Traffic#

A healthy community system is not just high chat volume or more guild joins. It is the creation of durable, satisfying, repeat social bonds.

4. System Proposal#

This is the foundational social data model above the raw friend list.

It should capture:

durable relationships: friend, guildmate, mentor, mentee, recurring squadmate, alliance partner, blocked, avoided
intent: questing, raiding, crafting, exploration, casual chat, role-play, PvP, teaching, learning
context: schedule windows, voice preferences, language, region, activity duration, platform
social comfort: prefers AI-first onboarding, no-mic groups, small groups, teaching-friendly groups, first-time comfort
reliability: show-up rate, disconnect rate, completion rate, endorsement rate
social trajectory: first group completed, first friend added, first recurring squad, first guild joined, mentor relationship formed

The graph should mix explicit player settings with observed behavior, but observed inference must remain inspectable and overridable by the player.

Key capabilities:

intent profiles players can edit quickly
ephemeral "what I want right now" state
recent positive co-play memory
relationship-strength scoring
privacy and visibility controls per field
soft compatibility features for discovery and routing

4.2 Discovery & Serendipity Layer#

Most multiplayer systems expose search. Stronger ones also create useful discovery.

The discovery layer should surface:

compatible players currently doing adjacent content
parties needing one more player for a matching quest line
guilds aligned to the player's schedule, role, and culture
recurring public gatherings the player is likely to enjoy
old squadmates who are active again
low-pressure events for shy or first-time players
"you almost formed a great team last night" follow-ups

This should appear in multiple places:

in-game home feed
quest journal
map/bulletin boards
town social hubs
companion web/mobile surfaces

The critical product requirement is that discovery should feel like useful social curation, not algorithmic spam.

4.3 Group Formation, LFG, and Quest Matching#

This layer extends beyond ordinary matchmaking and should cover both immediate and scheduled cooperation.

Core features:

structured LFG/LFM listings with intent tags
activity type, objective, tone, language, platform, voice, first-timer friendliness, teaching willingness, and role needs
now vs scheduled play
auto-join vs application
quest-specific recruitment directly from the quest journal
dungeon/raid/event listings with difficulty expectations
join-in-progress and backfill
re-queue with same group
post-run "stay together" prompts
saved squad templates and recurring time slots

The lesson from Bungie's Fireteam Finder and FFXIV's Party Finder is that players do better when expectations are structured up front.

Important additions for Oshun:

"quiet run" and "gentle pace" tags
"learning run" vs "speed clear" separation
"new here" and "returning player" lanes
AI-assisted listing composition for players who do not know what to write
expectation summaries shown before ready-up

4.4 Guild Recruitment & Community Finder#

Recruitment should not depend on chat spam or external Discord posts.

The system needs searchable guild/community profiles with:

playstyle focus
size
peak active hours
language
role needs
teaching culture
hardcore/casual positioning
PvE/PvP/role-play/crafting focus
voice expectations
cross-platform posture
new player friendliness
leadership activity
event cadence
open housing/guild hall availability

Required flows:

player search for guilds
guild search for players
public applications with short notes
recruiter review queues
provisional membership
auto-accept for compatible applicants when configured
recruiter office hours and calendar slots
open house listings
guild referrals from existing members

The core design goal is to help players find a social fit, not just a slot.

4.5 Guild Operations & Shared Progress#

Communities retain members when there is shared work and visible collective momentum.

Guild systems should support:

project boards
shared resource drives
weekly and seasonal goals
contribution tracking
role-based permissions
member onboarding queues
guild calendars and RSVP
announcement channels
shared storage and budget tools
knowledge base and playbook pages
roster notes and member strengths
recognition of invisible labor such as logistics, mentoring, and hosting

The EVE Corporation Projects model is especially instructive here: groups need a way to direct effort, recognize contribution, and lower ambiguity around how to help.

4.6 Alliances, Coalitions, and Cross-Community Cooperation#

A healthy MMO ecosystem cannot stop at single guilds.

Needed structures:

alliances of guilds
temporary coalitions around events or wars
guest access between guilds
shared project boards for alliance-scale goals
shared calendar surfaces
alliance chat and leadership channels
diplomatic relations and trust settings
inter-guild referrals
joint event hosting
shared newcomer pipelines

This prevents community fragmentation and lets smaller guilds remain viable by forming stable federations instead of dissolving into megaguilds.

4.7 New Player Onboarding, Returner Rescue, and Mentorship#

This is the most important community-growth layer after LFG.

The objective is not simply to teach controls. It is to shepherd the player toward their first stable social footholds.

Core capabilities:

newcomer identification
returner detection
guided first-session social goals
novice chat/help channels
mentor certification and routing
mentor availability windows
first-group recommendations
first-guild recommendations
first recurring teammate prompt
returner recap of what changed, who is still active, and where their old group went
"ask a guide" flows
low-pressure mentor escort runs

The system should explicitly optimize for:

time to first positive group experience
time to first friend added
time to first repeat teammate
time to first community membership

Mentor systems also need quality controls:

mentor reputation
response quality
abuse prevention
inactivity decay
protected novice cohorts

Community cannot live only in UI.

The world itself should create recurring opportunities for contact through:

starter taverns
city plazas
guild fairgrounds
quest boards
bounty boards
public training dummies
co-op crafting stations
music/performance spaces
fishing piers
campfires and caravans
seasonal festivals
faction ceremonies

These spaces should have social functions, not only decoration:

lightweight public activities
group formation kiosks
visible event schedules
ambient conversation prompts
reputation display
guild recruitment booths
mentor offices
AI/NPC hosts

Recurring world rituals are especially valuable because they create predictable co-presence. People form routines around repeated, familiar meetings.

4.9 Reputation, Trust, Reliability, and Compatibility#

Healthy community systems need more nuance than "report or friend."

Recommended layers:

positive endorsements
reliability scores
teaching friendliness
social fit markers
completion history
event attendance
mute/block/avoid
comfort preferences
probation states after misconduct

This system must be designed very carefully. Reputation is useful, but it can easily become biased, gamed, exclusionary, or humiliating.

Therefore:

use context-specific signals
privilege reliability and safety over popularity
avoid public global social scores
keep sensitive scores private and matchmaking-oriented
provide appeal/recovery pathways
use anti-brigading protections

4.10 Communication, Moderation, and Community Safety#

Communication systems in Phase 77 provide the baseline. This proposal adds the community-health layer around them.

Needed capabilities:

rules acceptance before participation
channel/role self-selection during onboarding
newcomer-safe channels
guild and alliance moderation tools
pre-post automated filtering
spam, scam, raid, and anomaly detection
moderator alert channels
evidence-rich reporting
case management
voice safety escalation
block, mute, ignore, and shadow routing tools
regional language moderation

The Riot Player Dynamics lens is important here: some toxicity is intentional, but a large amount of social friction is situational. Product design should remove common causes of avoidable conflict:

mismatched expectations
unclear group goals
forced contact under stress
no exit or regroup path
no place for quieter players
vague moderation rules

4.11 Leadership, Governance, and Community Operations#

Communities weaken when leadership tooling is primitive.

Guilds and public communities need:

officer roles with scoped permissions
event ownership
recruitment ownership
moderation delegation
succession planning
absence coverage
governance notes and decision logs
polls and lightweight voting
member segmentation
burnout monitoring for volunteer leaders
shared runbooks for welcoming, conflict resolution, and escalation

For official ecosystem communities, Oshun should also support:

ambassador programs
mentor councils
creator/community spotlights
community advisory groups

Blizzard's WoW Ambassador direction is notable here: trusted community members can become structured social guides rather than informal volunteers with no support.

4.12 Recognition, Status, and Prosocial Incentives#

People stay where effort is seen.

Useful recognition surfaces include:

mentor badges
reliable raid lead status
community builder titles
newcomer helper distinctions
guild host spotlights
public thank-you notes
anniversary markers
guild memory books
event founder recognition
alliance service honors

These should emphasize:

welcoming
teaching
reliability
contribution
hosting
conflict de-escalation

They should not create a manipulative social economy where players perform kindness for leaderboard rank alone.

4.13 External Community Surfaces & Companion Experiences#

Modern multiplayer community extends beyond the client.

The platform should expose companion surfaces for:

guild search
event calendars
recruitment applications
RSVP and reminders
mentor booking
community announcements
activity feeds
deep links into group listings
Discord/community server integration
web profile pages for guilds and events
returner briefings

This is where Lilith companion surfaces can help. Players should be able to:

discover groups while away from the game
schedule play in advance
reconnect with their community before logging in
receive socially meaningful notifications rather than generic churn spam

4.14 Community Analytics, Health Modeling, and Experimentation#

If Oshun wants to build a truly healthy MMO ecosystem, community health must be observable.

The analytics layer should track:

time to first group
time to first friend
time to first repeat teammate
time to first guild
guild conversion rate
guild retention
repeat squad formation
event attendance
mentor response rates
community churn
harassment incidence
moderation backlog
LFG abandonment
post-session reformation rate
returner reintegration rate

It should also support:

community-health dashboards
isolation-risk modeling
guild health scoring
leader tool adoption
A/B testing of social surfaces
safe experimentation frameworks

The goal is not just retention. The goal is to detect whether players are becoming more connected or more isolated.

This is the new layer requested explicitly: AI that helps players come out of their shell and meet people.

This system should be built around transparent, opt-in social facilitation.

Personal guide agent Helps a player understand the social landscape, recommends low-pressure groups, and explains etiquette or expectations.
Social wingman Acts like an outgoing friend who helps the player take the first step: suggests who to approach, drafts an introduction, encourages participation, and can accompany the player into the early moments of an interaction.
Group concierge Summarizes the party before the player joins: who is teaching, who is new, what the goal is, what tone to expect.
Returner rescue agent Helps a returning player understand what changed, which friends or guilds are still active, and what a good re-entry path looks like this week.
Mentor assistant Routes questions to the right mentors, drafts context summaries, and handles low-stakes repeated questions.
Companion fallback teammate Fills an empty slot temporarily so a player is not stranded, while still nudging toward eventual human group play.

For the exact use case the user described, the AI should be able to:

notice the player repeatedly hovers over a listing but does not join
offer a low-pressure explanation of the group
suggest a compatible first message
offer to announce the player's goal in a structured, disclosed way
introduce the player using concise shared context: "This player is doing this dungeon for the first time and prefers a calm run"
encourage the player to speak or emote at the right moment
help them follow up after a good run: "You played well together. Want to queue again or send a friend request?"
coach before and after interaction without making the player feel judged

The correct mental model is: AI as disclosed social scaffolding.

4.15.3 AI Rehearsal and Confidence Building#

Some players need a place to practice before joining real people.

The agent can provide:

mock party-join conversations
etiquette rehearsal
raid/guild introduction practice
explanation of group slang
low-stakes confidence prep

This is especially useful for:

new MMO players
younger players
socially anxious players
players joining voice for the first time
players switching regions/languages

4.15.4 AI in Live Group Play#

During live sessions, the AI can:

summarize unclear objectives
remind the player when to ask for clarification
suggest lightweight icebreakers
help quieter players contribute at natural moments
highlight positive behavior from other players
recommend staying with a group after a strong run

If deployed in voice or text spaces, the AI must always be visibly labeled and easily muted or dismissed.

4.15.5 AI Guardrails#

This is the most important part.

The AI must never:

pretend to be a human player
pressure a player into unwanted social contact
reveal private information or inferred vulnerabilities
create parasocial dependency by discouraging human connection
manipulate emotions to maximize engagement
continue intervening after the player opts out

Instead, it should:

disclose its role clearly
ask before stepping into socially meaningful interactions
keep interventions concise
prefer handoff to humans
respect comfort and privacy settings
support recovery after bad experiences

4.15.6 AI Technical Direction#

The right architecture here is hybrid:

lightweight, fast local/on-device models for immediate coaching, suggestion, and session summarization
server-side orchestration for memory, scheduling, social graph retrieval, and safety policy
structured tool use against community services rather than free-form improvisation alone

This should integrate with:

social graph
LFG listings
guild recruitment
mentor network
event calendar
trust and safety systems

World elements and NPCs can also help create human connection.

Examples:

tavern keepers that surface nearby groups seeking one more player
faction officers that recommend guilds aligned to the player's goals
bulletin board stewards that translate quest logs into joinable public tasks
caravan masters that create shared travel missions
public event hosts that repeatedly gather the same categories of players
quiet-space NPCs that direct socially cautious players to smaller or calmer gatherings
festival MCs that highlight mentor events, guild fairs, and public activities

These NPCs should not replace player communication. They should route, host, and reduce ambiguity.

World choreography can further increase social formation through:

recurring event timing
public crafting and repair bottlenecks
co-op mini-objectives in safe zones
proximity introductions around campfires or staging grounds
ambient "looking for help" signals in settlements
rotating neighborhood/guild fair events

The important principle is simple:

Social encounters should be designed into the geography and rhythms of the world.

5. Architecture Recommendations#

5.1 Proposed Service and Library Layer#

Recommended additions across the Oshun stack:

libs/maya/social-graph/
libs/maya/social-discovery/
libs/maya/lfg-community/
libs/maya/guild-recruitment/
libs/maya/guild-operations/
libs/maya/mentorship/
libs/maya/community-health/
libs/maya/trust-safety/
libs/maya/community-events/
libs/maya/social-facilitation/
libs/maya/community-companion/
libs/maya/community-governance/
libs/maya/community-reputation/
libs/maya/world-social-runtime/

Runtime/world integration:

Rust crates in libs/maya/engine-core/crates/ for NPC hosts, world bulletin boards, social hubs, and event choreography

Companion/admin surfaces:

Lilith companion pages for recruitment, calendar, mentorship, and social health operations

Analytics:

Sophia pipelines for health scoring, experimentation, and insight generation

5.2 Core Data Models#

Suggested entities:

PlayerSocialProfile
PlayerIntentState
AvailabilityWindow
GroupListing
QuestRecruitmentListing
GuildRecruitmentProfile
GuildProject
MentorProfile
MentorshipMatch
SocialReputationEvent
ReliabilitySnapshot
CommunityHealthSnapshot
SocialFacilitationSession
NpcSocialHostDefinition
WorldSocialEvent

5.3 Integration Points with Existing Phases#

Phase 77: friends, parties, guilds, chat, events, lobbies, cross-platform identity
Phase 79: living world systems, crowd behavior, NPC social simulation, festivals, urban spaces
Lilith: companion/community web and mobile surfaces
Sophia: research, analytics, cohort modeling
Hathor: narrative framing for social rituals, festivals, mentorship lore, faction identity

6. Implementation Priorities#

Build first:

intent graph
structured LFG/community listings
guild/community finder
newcomer and returner flows
mentor network
post-session reformation
companion notifications

This is the highest-impact layer for actual community formation.

Priority 2: Community Health & Guild Lifecycle#

guild operations and project boards
alliance features
reputation and reliability
safety onboarding
moderation tooling
leadership tools
analytics dashboards

Then:

AI guide
AI wingman
mentor assistant
returner rescue
companion fallback teammate
group summaries and coaching

Finally:

tavern hosts
bulletin board orchestration
public ritual design
festival routing
world social spaces and event cadence

7. Success Metrics#

The top-level KPI stack should include:

time to first group
time to first positive repeat group
time to first friend
time to first guild/community
7-day and 30-day newcomer social retention
percentage of players with at least one recurring teammate
percent of guild joins that remain active after 14 days
mentor response SLA
returner reintegration rate
LFG conversion rate
post-run "stay together" conversion
safety incident rate per social surface
moderation time to action
share of players using low-pressure/shy-player features
AI wingman opt-in satisfaction

The north-star metric should be:

Percentage of active players with at least one durable, positive recurring human connection in the ecosystem.

8. Risks and Anti-Patterns#

Having guilds, chat, and parties is not the same as having community.

8.2 Over-Optimizing for High-Skill Players#

A community stack that works only for confident, vocal, or hardcore players will bleed everyone else.

8.3 Public Reputation Systems that Become Exclusion Engines#

Trust signals should reduce harm, not create permanent social castes.

8.4 AI Deception#

If AI pretends to be human or quietly manipulates player behavior, trust in the entire ecosystem will collapse.

Some players want gradual engagement. Systems should support solitude and agency, not pathologize them.

8.6 Measuring Volume Instead of Belonging#

More messages or more listings do not necessarily mean stronger communities.

8.7 Cliques Hardening into Closed Networks#

Recommendation systems should preserve access for newcomers and avoid endlessly recycling the same small social graph.

9. Reference Links#

Bungie Fireteam Finder Guide: https://help.bungie.net/hc/en-us/articles/25787853699220-Fireteam-Finder-Guide
Bungie Destiny 2 Clans Guide: https://help.bungie.net/hc/en-us/articles/360048720932-Destiny-2-Clans-Guide
FFXIV Party Play / Duty Finder / Party Finder: https://na.finalfantasyxiv.com/game_manual/pp/
FFXIV Free Company overview: https://na.finalfantasyxiv.com/lodestone/special/grandcompany/about_freecompany/
FFXIV Free Company Search: https://na.finalfantasyxiv.com/lodestone/freecompany/
FFXIV community overview: https://na.finalfantasyxiv.com/a_realm_reborn/system/community/
WoW Communities: https://worldofwarcraft.blizzard.com/en-us/news/21952590/battle-for-azeroth-preview-world-of-warcraft-communities
WoW Guild & Communities Finder: https://worldofwarcraft.blizzard.com/en-us/news/23232341
WoW Ambassador Program: https://worldofwarcraft.blizzard.com/en-us/news/24244402
Blizzard support for WoW Communities: https://us.support.blizzard.com/article/000311333
WoW Cross-Faction Guilds: https://worldofwarcraft.blizzard.com/en-us/news/23934662/play-with-more-friends-with-cross-faction-guilds
EVE Corporation Projects: https://support.eveonline.com/hc/en-us/articles/9583433729308-Corporation-Projects
EVE Freelance Jobs: https://support.eveonline.com/hc/en-us/articles/20210930509340-Freelance-Jobs
EVE AIR Career Program: https://support.eveonline.com/hc/en-us/articles/5194673685148-AIR-Career-Program
EVE Access Lists: https://support.eveonline.com/hc/en-us/articles/208289645-Access-Lists
EVE Legion overview: https://www.eveonline.com/news/view/legion-in-focus
Warframe Clan, Alliance and Dojo Guide: https://support.warframe.com/hc/en-us/articles/200049929-Clan-Alliance-and-Dojo-Guide
Warframe Cross-Platform Friends & Clans/Alliance: https://support.warframe.com/hc/en-us/articles/20686206963213-Cross-Platform-Friends-Clans-Alliance
Discord Enabling Your Community Server: https://support.discord.com/hc/en-us/articles/360047132851-Enabling-Your-Community-Server
Discord Rules Screening FAQ: https://support.discord.com/hc/en-us/articles/1500000466882-Rules-Screening-FAQ
Discord AutoMod FAQ: https://support.discord.com/hc/hi-in/articles/4421269296535-AutoMod-FAQ
Discord Server Insights FAQ: https://support.discord.com/hc/en-us/articles/360032807371-Server-Insights-FAQ
PlayFab Community overview: https://learn.microsoft.com/en-us/gaming/playfab/community/what-is-community
PlayFab Entity Groups: https://learn.microsoft.com/en-us/gaming/playfab/community/associations/groups/quickstart
PlayFab searchable lobbies: https://learn.microsoft.com/en-us/gaming/playfab/multiplayer/lobby/define-search-keywords
PlayFab lobby + matchmaking: https://learn.microsoft.com/en-us/gaming/playfab/features/multiplayer/lobby/lobby-and-matchmaking
Xbox Social Manager overview: https://learn.microsoft.com/en-us/gaming/gdk/docs/services/community/social-manager/live-social-manager-overview
Riot Player Dynamics: https://www.riotgames.com/en/news/player-dynamics-design-looking-behind-the-curtain
NVIDIA ACE autonomous characters: https://www.nvidia.com/en-us/geforce/news/nvidia-ace-autonomous-ai-companions-pubg-naraka-bladepoint/
Generative Agents paper: https://arxiv.org/abs/2304.03442