LOBBY → ROLE_SELECTION → IN_PROGRESS → STAGE_1 → STAGE_2 → STAGE_3 → COMPLETED

Stage → Modules → Scenarios → Player Responses → AI Injection → Consensus → Next

Conversational Game Project Plan

Overview

This project aims to create a web-based, AI-assisted multiplayer conversational game with executive and technical tracks. Users can host or join games, assume roles, interact with AI agents, and progress through multiple stages. A final report evaluates performance and collaboration.

1. User Workflow

1. Sign Up / Login

   • Users create an account or log in.
   • Basic profile information is stored (username, email).

2. Select Track

   • Users choose between:
     • Executive Track
     • Technical Track

3. Host or Join Game

   • Host: Generates a unique invite code.
   • Join: Enters invite code to join an existing game.

4. Role Selection

   • Players select roles (CEO, CISO, Legal, HR, PSIRT Analyst, Security Manager, Facilitator).
   • Unfilled roles are automatically assigned to AI agents.

5. Game Start

   • Game begins when all roles are filled or AI agents are assigned.
   • Game progresses through 3 stages, each containing multiple modules/scenarios.

2. Game Flow

1. Stages and Modules

   • Each stage has x modules.
   • Modules are challenging scenarios where players must provide inputs.
   • Facilitator AI injects additional complexity or risks to simulate realistic pressure.

2. Player Interaction

   • Players submit decisions or responses for each module.
   • AI agents provide role responses if roles are unfilled.
   • One player may act as a saboteur, intentionally creating challenges for other players.

3. Stage Progression

   • Consensus or completion of module triggers next module.
   • Completion of all modules in a stage triggers next stage.
   • After Stage 3, the game ends.

3. AI Integration

1. Facilitator AI

   • Injects complexity and challenges.
   • Guides players but does not give direct answers.

2. Role AI

   • Fills unassigned player roles.
   • Responds according to role objectives and personality.

3. Saboteur AI (Optional)

   • Introduces subtle conflicts or delays.
   • Challenges proposals to simulate real-world adversities.

4. Evaluation AI

   • Reviews actions from all players and AI agents.
   • Generates a performance report highlighting strengths, weaknesses, and recommendations.

4. Technical Architecture (MVP)

• Frontend: Bubble / Glide / Adalo (No-Code)

  • Pages: Login/Signup → Lobby → Role Selection → Game Interface → Report
  • Components: Player cards, Module display, Chat window, Progress tracker

• Backend / Database: Airtable or Bubble DB

  • Tables: Users, Games, Players, Stages, Modules, Actions

• AI Integration: OpenAI API via Make.com / Zapier

  • AI agents simulate roles and facilitator interactions.

• Multiplayer Logic:

  • MVP can be turn-based submission instead of real-time.
  • Later, real-time can be added via Firebase or WebSockets.

5. Project Phases & Milestones

Phase 1: MVP Setup

• Setup user authentication and accounts
• Host/Join game flow
• Role selection (human + AI)
• Basic stage/module navigation

Phase 2: Core Gameplay

• Module scenarios implemented
• Facilitator and role AI integration
• Turn-based player interactions

Phase 3: AI Complexity

• Saboteur AI (optional)
• Dynamic facilitator injections
• Conditional scenario branching

Phase 4: Reporting & Analytics

• Post-game evaluation
• Individual and team performance scoring
• Recommendations generated by AI

Phase 5: Polish & Deployment

• UI enhancements
• Multi-device support
• Beta testing with users
• Optional: Real-time multiplayer upgrade

6. Notes

• Keep MVP simple to launch quickly.
• Focus on game flow, role dynamics, and AI responses.
• Real-time multiplayer can be added later.
• AI prompts should be structured JSON for consistent game state updates.

Step 1: Pick No-Code Platforms

You need platforms for frontend, backend, and AI:

✅ Why Bubble:

• Handles login/signup

• Lets you design pages visually

• Can create “lobbies” and role selection pages

• Can call external APIs (like OpenAI)

• Has a built-in database

Step 2: Design Your Pages in Bubble

1. Login / Signup Page

   • Fields: Username, Email, Password

   • Button: “Sign Up” / “Login”

2. Lobby Page

   • Buttons: “Host Game” → generates invite code

   • “Join Game” → input invite code

   • Track selection dropdown: Executive / Technical

3. Role Selection Page

   • List of roles: CEO, CISO, Legal, HR, Facilitator, etc.

   • Players select roles; unselected roles are assigned to AI automatically

4. Game Page

   • Module display: Scenario text

   • Input box: Player submits actions/decisions

   • Chat box: Shows responses from other players and AI agents

   • Progress bar: Shows stages/modules

5. Report Page

   • Shows individual and team performance

   • Highlights good decisions, mistakes, and suggestions

Step 3: Set Up the Database

Tables / Data Types in Bubble DB:

• Users: Username, Email, Password

• Games: Host, Invite Code, Track, Status (Waiting / In Progress / Completed)

• Players: Game, User, Role, Is AI

• Stages / Modules: Game, Stage Number, Module Number, Scenario Text, Status

• Actions: Module, Player, Input, Timestamp

Step 4: Game Logic (Simplified for MVP)

Since we avoid coding, your MVP should be turn-based, not fully real-time:

1. Players submit actions for the module

2. AI responds to unfilled roles and facilitator messages

3. Once all players (or AI) have submitted → next module appears

4. After all stages/modules → show final report

✅ Tip: Bubble’s workflows can handle this “if all actions submitted → move to next module” logic visually.

Step 5: AI Integration

You can integrate OpenAI API via Make.com (or Zapier):

1. Trigger: Player submits action

2. Action: Call OpenAI API with a prompt for the AI agent

   • Example: Facilitator AI → “Inject complexity in current scenario”

   • Example: Role AI → “Respond as Legal Officer to player decisions”

3. Return AI response → show in chat box on game page

Structured Prompt Example for OpenAI:

{
“role”: “Legal”,
“scenario”: “Player has decided to disclose a data breach publicly”,
“objective”: “Ensure compliance and reduce liability”,
“response_format”: {
“action”: “text”
}
}

Step 6: MVP Launch Steps

1. Build pages in Bubble

2. Connect database tables (users, games, players, modules)

3. Design workflows:

   • Signup/Login

   • Host/Join Game

   • Role selection

   • Submit module action → AI response → move to next module

4. Integrate OpenAI via Make.com or Zapier for AI agents

5. Test the game internally with friends or colleagues

6. Show the final report page with AI-generated evaluations

Step 7: Keep It Simple

For the MVP, don’t worry about:

• Real-time updates (use turn-based module submissions)

• Complex UI/animations

• Multiple concurrent games (one at a time is fine)

• Full AI sophistication (start with facilitator + 1 AI role)

The goal is a playable prototype that demonstrates your idea.

If you want, I can make a “step-by-step Bubble MVP blueprint” with exact page layouts, database tables, and AI integration instructions specifically for your conversational game.

Do you want me to make that next?

i also want to make it modular, like i can create more scenarios etc

Perfect! Making it modular is smart—it lets you add new scenarios, stages, or roles without touching the core game logic. Since you’re using a no-code approach like Bubble, modularity mostly comes from database design + workflows. Here’s how to do it:

1️⃣ Database Design for Modular Gameplay

Games

• id (unique)

• host (user)

• track (executive / technical)

• invite_code

• status (waiting / in_progress / completed)

Stages

• id

• game_id (link to game)

• stage_number

• name

• status (pending / in_progress / completed)

Modules (Scenarios)

• id
• stage_id (link to stage)
• module_number
• scenario_text
• type (decision / challenge / discussion / simulation)
• status (pending / completed)
• optional: facilitator_injection_text

Roles

• id
• name (CEO, CISO, Legal…)
• objectives (JSON/text)
• AI_agent_prompt (for AI to simulate this role)

Players

• id
• game_id
• user_id (nullable if AI)
• role_id
• is_ai

Actions

• id
• module_id
• player_id
• input_text
• timestamp

✅ This lets you create new scenarios/modules just by adding records to the “Modules” table. No workflow changes needed.

2️⃣ Modular Game Flow

1. Stages → Modules

   • Each stage can have any number of modules.
   • Bubble workflow: “Go to next module” → dynamically pulls the next module record from the database.

2. Roles

   • Adding new roles = adding a new record to the Roles table.
   • AI automatically picks up prompts from AI_agent_prompt.

3. Facilitator injections

   • Each module can have an optional field for “complexity” or “injection”.
   • Bubble workflow → AI reads this and outputs messages.

4. AI Responses

   • AI prompts use the module text + role objectives → returns JSON.
   • Since prompts are stored per role/module, adding a new scenario is just a database insert, not workflow redesign.

3️⃣ How to Add Scenarios / Modules Easily

• Admin page in Bubble:
  • Add stage (name, order)
  • Add modules to that stage (scenario text, type, facilitator injection)
  • Assign roles that are expected to participate
• Optional: Tag modules with difficulty or category so you can filter or randomize scenarios per game.

4️⃣ AI Prompts for Modularity

Store prompts per role and per scenario type, for example:

{
“role”: “Legal”,
“scenario_type”: “Data Breach”,
“objective”: “Ensure compliance and reduce liability”,
“instructions”: “Respond in a professional tone. Provide guidance, highlight risks, do not solve directly.”
}

• When a new scenario is added → workflow fetches the relevant AI prompt automatically.
• This makes AI behavior consistent across scenarios without coding.

5️⃣ Recommended Workflow in Bubble for Modular MVP

1. Game Start → Fetch stages from DB → Load Stage 1
2. Module Load → Fetch module from DB → Display scenario
3. Player Submission → Save action to DB → Trigger AI agents for unfilled roles
4. Facilitator Injection → Optional AI message → Display
5. Module Completion → If all players/AI submitted → Move to next module
6. Stage Completion → If all modules done → Move to next stage
7. Game Completion → Generate report (pull data from Actions table, evaluate, show summary)

✅ Adding a new scenario or stage = just adding records in DB.

If you want, I can now make a complete Bubble MVP blueprint that includes:

• Page layout
• Tables (DB) setup
• Admin page for adding stages/modules
• Modular AI integration flow

It will literally be a step-by-step guide you can follow without coding.

Do you want me to create that next?