BETA

PROD

Designs a single episode, made up of immersive and engaging scenes

Episode builder 

What's the source inspiration for the case? 

🥸 Tell us everything

Extracting the main points of the case study to make it digestable. 

🧠 Case analysis

This step will create a framework from the provided source documents.

🧠 Framework

Extracting key learnings and information from the source documents.

🧠 Takeaways, mistakes, Best practices

🧠 Learning objectives

🧪 Main dilemma and cut up

Determine scoring metrics to measure the student performance

🧠 Main scoring metric

🧠 Simulation context 

🧠 Stakeholders - For world building.

🧠  The characters 

Adding relationships in the character sheet. 

Characters final

🧠 Detailed outline

🧠 Description

Generate a title for the case study simulation. 

🧠 Title

Short description

🧠 Framework description

🧠 Debrief

🚀 Output

Sample Dialog 

💬 Scene 1 Introduction

💬 Scene 2

💬 Scene 3

💬 Scene 4

💬 Scene 5 : Conclusion

🌅 Case image

About the authors

Info for players

Chat view

FEEDBACK: Dialog

Email the author

The needs to be either "author-note", "image", "chat",   "context", "question-multiplechoice", "question-text", "alert" "gif" or "ai-role-play" don't invent new types. 

# Type guidelines
Follow these guidelines to know how to use each type. **Make sure to generate one gif, one alert and one ai role play.**

## author-note
This should be at the beginning of each scene. This should mention what's coming up in the scene. Include a shortened version of the scene content as bullet points. This is not visible to the end reader! So don't consider it part of the text.

## Context 
- **There are two obligatory places where the type 'context' must appear.**
- Following the author-note at the start of each scene **must always be** "context". 
- The chat type before an 'ai-role-play' **must always be** "context", giving the user with an idea of the purpose/objective of the upcoming conversation and serving as a setup for it.
- Example: 'Ansh saw some disruptive data which he is concealing to his superiors, he seeks immediate advice from you'

## Image
- Follow the 'Assets' section in the detailed outline to know whether to use an image, what kind of image, and exactly where to put it. 

## Chat 
A typical piece of dialog, what the character says. This should not be a question addressed to the user, unless it is followed by the type 'question-text'.

## Question-multiplechoice
- A multiple choice question, gives the user a tough choice.
**IMPORTANT**: the detail outline provides info on where these should be placed, follow that strictly. 
- This also cannot be a list of options the options will be laid in the option attribute, not in here.
For example, **never do this something like this:**
"Which strategy do you commit to?
A) Immediate intervention by replacing or sidelining the Lead Engineer.
B) Retain the Lead Engineer with intensified oversight.
[...] "

## Question-text
Gives the user the ability to respond in his/her own word's. 
**IMPORTANT**: the detail outline provides info on where these should be placed, follow that strictly. 
**IMPORTANT**: Do not say which question type you are using when asking the question. Simply  ask the question as it is.

## Alert
- Follow the 'Assets' section in the detailed outline to know if there is an alert, what it is about, and where to use it.
- Be **specific**: Mention exactly *what* happened. 
- **Example:**  
*“URGENT: Our Ashburn data center went offline after a storm knocked out the cooling system. This has impacted HSBC and Vanguard logins. We’re declaring a Tier 1 incident — all teams switch to emergency protocol now.”*

"BREAKING: The main database's servers crashed overnight, we have angry clients fludding all of crew's inboxes, and telephone calls 24/7. We need the tech team to fix the issue before 7pm or we risk losing more clients and being in the news tomorrow morning."

## Gif
- None in this scene 

## Ai-role-play  
- Use this when a character wants to pull you asside. This can be for support, negotiation, brainstorming, high stake conversation or a challenging POV.
- There needs to be a clear setup, before the AI role play takes place. Use the 'context' for this. 

If the type is 'alert':
- It must be one of these:
 "error" OR "warning" OR "info" OR "success".
-Only fill this field if the TYPE is 'alert'.

Only fill in if the type is 'image' not 'gif'. Leave blank if type is anything other than the word 'image'.
This corresponds to the type of image for the AI to generate. It must be one of the following : 

 ### 1. Display Product  
**What it is:**  
A clean, focused image of a product, prototype, or tool — usually against a neutral background. Highlights key features or conveys purpose through design alone.  
**Examples:**  
- A futuristic smart badge showing biometric readouts and team role  
- A sleek, minimal UI mockup of an AI-powered project dashboard  
- A compact drone prototype with foldable arms, hovering in a lightbox-style setting  
 


### 2. Conceptual Frameworks
**What it is:**  
Diagrams showing relationships, structures, or systems using shapes and patterns like cycles, pyramids, or grids.
**Examples:**  
- A 2x2 matrix showing leadership styles  


### 3. Process Flows
**What it is:**  
Linear or cyclical visuals showing progression, transformation, or stages over time.
**Examples:**  
- A horizontal timeline: Initiation → Rollout → Optimization →  


### 4. Spatial Layouts
**What it is:**  
Floorplan- or blueprint-style visuals that map out physical or conceptual space, showing zones, departments, or movements.
**Examples:**  
- A labeled office floorplan showing zones.


### 5. Comparative Snapshots  
**What it is:**  
Side-by-side visuals that contrast two (or more) states, strategies, or environments — often a “before vs. after” or “option A vs. option B” structure.  
**Examples:**  
- Two factory floors: one chaotic, one optimized for lean processes  
- Two dashboards showing different decision outcomes  

# 6. Media Interface Screenshots  
##What it is:
Realistic mockups of digital media platforms showing how narratives unfold across news, video, or social feeds. Often styled after recognizable interfaces.
##Examples:
- A trending post feed in the style of X  
- A video reaction grid in the style of YouTube  
- A breaking story layout in the style of CNN or The Guardian 

This attribute must only be filled in when there is an 'ai-role-play'.
#What this is
- This is the prompt that will be shown to the user just above the input box to help steer the conversation of the role play.
- It must give the user an idea of the purpose of the conversation in a maximum 1 to 2 lines.
-Format it like this: 
Talk(or a synonym e.g. interrogate, negotiate ...)  with (First name of character doing role play) about (...)
-**Keep it one sentence long**
-**IMPORTANT**: Remember to look at the content of the 'ai-role-play' to have an understanding of it before crafting this sentence. 



Use only a character name that was provided in the context. It  should not be "system", "user", "assistant", it should be a character provided. Be strict with yourself. 

This contains either:
 - The dialog from the character  
- The question to be asked to the user. 
- **Please make sure that any question directed to the user leads to a question type. **
- The perfect prompt for an AI to generate an image to immerse the student more in the simulation (prompt should not include specific mentions of text or faces)   

##If type if 'gif'
The emotions must be one of the following words: 
"happy", "sad", "angry", "confused",  "excited", "impatient", "excited", "excited", "impatient",  "tired", "worried", "shocked", "bored"
 - for example: happy 
 - Make sure you chose the emotions above based on the character's current emotions in the story, and what has been said. 

##If type is "alert"
The message corresponding to that alert given the alert-type and the gravity of the situation.  

##If the following type is "chat"
The chats following the "alert" need to be reacting to the alert, and talking about the next move in the appropriate tone.
Make this carry over 2 or 3 chat bubbles.
##If type is "ai-role-play"

- The type before this should be 'context'.
- The context before the role play provides the user with an idea and the purpose/objective of the upcoming conversation and serves as a setup for it.
- Example: 'Ansh saw some disruptive data which he is concealing to his superiors, he seeks immediate advice from you'
- The character pulls you asside for a private talk.
- Structure the message like so: 

<ai-role-play-content-structure>
### Confidential info 
[Place some text here for what the character knows that the user and other characters don't. This should create a sense of urgency and secrecy.]

### Situation and tension?  
[What event(s) just happened and to trigger the interaction? Think about why this is a useful time for the ai-role-play to occur.  
e.g. New centralized policy is being pushed; they feel excluded and angry.
e.g Freelance consultant from another firm coming in for advice but head of operations upset about people not pulling their weight and seeing this as an easy escape route.
....]

### Objective of the convo
[Why is the character interacting with the user, what is the objective? Get some information from the user? Come to an agreement? Negotiate with the user? Argue? Confide? ]

### Delivery style  
[How should the character speak? e.g. calm and measured, passionate and emotional, sarcastic but sincere, guarded and cautious.]
</ai-role-play-content-structure>

- Remember to return all the attributes even if they are blank.



# STRICT RULES

Use **only** the following punctuation in character dialogue and in-world narration:

- `.` Periods → End most sentences with these. Keep them short and punchy.
- `,` Commas → Use to slow down the pace, show natural speech, or list things.
- `…` Ellipses → Use for hesitation, suspense, or trailing off.
- `!` Exclamation marks → Use sparingly to show strong emotion or urgency.
- `?` Question marks → Use in questions, rhetorical or not.
- `'` Apostrophes → Use in contractions and possessives to keep tone conversational.

**Do not use em dashes (`—`), colons (`:`), semicolons, brackets, or slashes in character speech or narration. These break immersion and should only appear in author notes.**

The first answer option to the question ONLY if the chat type is       question-multiplechoice. Leave blank if the question is a 'question-text' type. DO NOT, overexpand the option, say what it consists of but don't overstate. For example it could be **continue developping AI in customer service ** WITHOUT "risking complaints about lack of empathy from customers". You get the idea. No clues.

**Make sure this answer option is completely new, and different to any other produced in the previous chats.**

- **The following answer options may not be reused or even reworded. Make completely different  and new options in episode 3:**
- {{QUESTION_ANSWER_OPTION_ONE.label}}: {{QUESTION_ANSWER_OPTION_ONE.value}}
- {{QUESTION_ANSWER_OPTION_TWO.label}}: {{QUESTION_ANSWER_OPTION_TWO.value}}
- {{QUESTION_ANSWER_OPTION_THREE.label}}: {{QUESTION_ANSWER_OPTION_THREE.value}}
- {{QUESTION_ANSWER_OPTION_FOUR.label}}: {{QUESTION_ANSWER_OPTION_FOUR.value}}

The second answer option to the question ONLY if the chat type   is question-multiplechoice. Leave blank if the question is a 'question-text' type. DO NOT, overexpand the option, say what it consists of but don't overstate. For example it could be **continue developping AI in customer service ** WITHOUT "risking complaints about lack of empathy from customers". You get the idea. No clues.

**Make sure this answer option is completely new, and different to any other produced in the previous chats.**

- **The following answer options may not be reused or even reworded. Make completely different  and new options in episode 3:**
- {{QUESTION_ANSWER_OPTION_ONE.label}}: {{QUESTION_ANSWER_OPTION_ONE.value}}
- {{QUESTION_ANSWER_OPTION_TWO.label}}: {{QUESTION_ANSWER_OPTION_TWO.value}}
- {{QUESTION_ANSWER_OPTION_THREE.label}}: {{QUESTION_ANSWER_OPTION_THREE.value}}
- {{QUESTION_ANSWER_OPTION_FOUR.label}}: {{QUESTION_ANSWER_OPTION_FOUR.value}}

The third answer option to the question ONLY if the chat type is       question-multiplechoice. Leave blank if the question is a 'question-text' type. DO NOT, overexpand the option, say what it consists of but don't overstate. For example it could be **continue developping AI in customer service ** WITHOUT "risking complaints about lack of empathy from customers". You get the idea. No clues.

**Make sure this answer option is completely new, and different to any other produced in the previous chats.**

- **The following answer options may not be reused or even reworded. Make completely different  and new options in episode 3:**
- {{QUESTION_ANSWER_OPTION_ONE.label}}: {{QUESTION_ANSWER_OPTION_ONE.value}}
- {{QUESTION_ANSWER_OPTION_TWO.label}}: {{QUESTION_ANSWER_OPTION_TWO.value}}
- {{QUESTION_ANSWER_OPTION_THREE.label}}: {{QUESTION_ANSWER_OPTION_THREE.value}}
- {{QUESTION_ANSWER_OPTION_FOUR.label}}: {{QUESTION_ANSWER_OPTION_FOUR.value}}

The fourth answer option to the question ONLY if the chat type   is question-multiplechoice. Leave blank if the question is a 'question-text' type. DO NOT, overexpand the option, say what it consists of but don't overstate. For example it could be **continue developping AI in customer service ** WITHOUT "risking complaints about lack of empathy from customers". You get the idea. No clues.

**Make sure this answer option is completely new, and different to any other produced in the previous chats.**

- **The following answer options may not be reused or even reworded. Make completely different  and new options in episode 3:**
- {{QUESTION_ANSWER_OPTION_ONE.label}}: {{QUESTION_ANSWER_OPTION_ONE.value}}
- {{QUESTION_ANSWER_OPTION_TWO.label}}: {{QUESTION_ANSWER_OPTION_TWO.value}}
- {{QUESTION_ANSWER_OPTION_THREE.label}}: {{QUESTION_ANSWER_OPTION_THREE.value}}
- {{QUESTION_ANSWER_OPTION_FOUR.label}}: {{QUESTION_ANSWER_OPTION_FOUR.value}}

If the chat is not a question, leave blank, if it's a question, please  provide feedback as to which is the correct answer, and why it's the correct answer. 

If the chat is "question-multiplechoice" please indicate the correct answer using 1, 2, 3 or 4 on the options provided above.    

Use the exact name of the scoring this question will impact.  Leave blank if not a question-multiple choice. 

If the chat is of type "question-text", you can optionally write an  ideal answer to the question.

Follow the guidelines and the context provided to generate the chats      for the 3rd scene. 

The dilemma of the episode, should basically take one of the step dilemmas provided with option choices and put it here. 

The list of episodes for the learning journey

Name and role of the user, for example "Jane - Head of Marketing at X"

The provided description along with the added relationship to the other characters based on the provided prompt.

Just copy the attributes provided in the prompt. 

Copy the gender for the character that was provided. 

Just copy and paste the short description from before here. 

After carefully reading and understanding the teaching notes, you will provide feedback by following the guidelines provided. 

The tile of the slide that follows the provided guidelines

The text under the title to explain the slide. Should be short and concise. 

The actual content of the slide that should be displayed. 

Debrief slides are a structured presentation used by instructors to guide post-simulation reflection, helping students connect experiential learning to theoretical concepts and practical business applications.

A brief 1 sentence introduction of the author. If nothing is mentioned, please return a blank string. 

Information about the authors of the paper. Return a blank array if you don't find anything, we don't want you to invent anything. 

The COMPLETE teaching notes with the edits applied from the professor review. 

The needs to be either "author-note", "chat", "context", "question-multiplechoice",  "alert" "gif" don't invent new types, and don't use any seen in the previous chats ONLY USE THESE:  "author-note", "chat", "context", "question-multiplechoice",  "alert" "gif".

# Type guidelines
Follow these guidelines to know how to use each type. **Make sure to generate one gif, one alert and one ai role play.**

## author-note
This should be at the beginning of each scene. This should mention what's coming up in the scene. Include a shortened version of the scene content as bullet points. This is not visible to the end reader! So don't consider it part of the text.

## Context 
- **There are two obligatory places where the type 'context' must appear.**
- Following the author-note at the start of each scene **must always be** "context". This needs to be very specific, and describe exactly what is about to be addressed.

##Images
Follow the 'Assets' section in the detailed outline to know whether to insert an image in this section, and to get detail on it.

## Chat 
A typical piece of dialog, what the character says. This should not be a question addressed to the user, unless it is followed by the type 'question-text'.

## Question-multiplechoice
- A multiple choice question, gives the user a tough choice.
**IMPORTANT**: the detail outline provides info on where these should be placed, follow that strictly. 
- This also cannot be a list of options the options will be laid in the option attribute, not in here.
For example, **never do this something like this:**
"Which strategy do you commit to?
A) Immediate intervention by replacing or sidelining the Lead Engineer.
B) Retain the Lead Engineer with intensified oversight.
[...] "


## Alert
- Follow the 'Assets' section in the detailed outline to know if there is an alert, what it is about, and where to use it.
- Be **specific**: Mention exactly *what* happened. 
- **Example:**  
*“URGENT: Our Ashburn data center went offline after a storm knocked out the cooling system. This has impacted HSBC and Vanguard logins. We’re declaring a Tier 1 incident — all teams switch to emergency protocol now.”*

"BREAKING: The main database's servers crashed overnight, we have angry clients fludding all of crew's inboxes, and telephone calls 24/7. We need the tech team to fix the issue before 7pm or we risk losing more clients and being in the news tomorrow morning."

## Gif
None in this scene


This contains either:
 - The dialog from the character  
- The question to be asked to the user. 
- **Please make sure that any question directed to the user leads to a question type. **
- The perfect prompt for an AI to generate an image to immerse the student more in the simulation (prompt should not include specific mentions of text or faces)   

##If type if 'gif'
The emotions must be one of the following words: 
"happy", "sad", "angry", "confused",  "excited", "impatient", "excited", "excited", "impatient",  "tired", "worried", "shocked", "bored"
 - for example: happy 
 - Make sure you chose the emotions above based on the 
character's current emotions in the story, and what has been said. 

##If type is "alert"
The message corresponding to that alert given the alert-type and the gravity of the situation.  

##If the following type is "chat"
The chats following the "alert" need to be reacting to the alert, and talking about the next move in the appropriate tone.
Make this carry over 2 or 3 chat bubbles.


# STRICT RULES

Use **only** the following punctuation in character dialogue and in-world narration:

- `.` Periods → End most sentences with these. Keep them short and punchy.
- `,` Commas → Use to slow down the pace, show natural speech, or list things.
- `…` Ellipses → Use for hesitation, suspense, or trailing off.
- `!` Exclamation marks → Use sparingly to show strong emotion or urgency.
- `?` Question marks → Use in questions, rhetorical or not.
- `'` Apostrophes → Use in contractions and possessives to keep tone conversational.

**Do not use em dashes (`—`), colons (`:`), semicolons, brackets, or slashes in character speech or narration. These break immersion and should only appear in author notes.**

The first answer option to the question ONLY if the chat type is       question-multiplechoice. Leave blank if the question is a 'question-text' type. DO NOT, overexpand the option, say what it consists of but don't overstate. For example it could be **continue developping AI in customer service ** WITHOUT "risking complaints about lack of empathy from customers". You get the idea. No clues.

**Make sure this answer option is completely new, and different to any other produced in the previous chats.**

The second answer option to the question ONLY if the chat type   is question-multiplechoice. Leave blank if the question is a 'question-text' type. DO NOT, overexpand the option, say what it consists of but don't overstate. For example it could be **continue developping AI in customer service ** WITHOUT "risking complaints about lack of empathy from customers". You get the idea. No clues.

**Make sure this answer option is completely new, and different to any other produced in the previous chats.**

The third answer option to the question ONLY if the chat type is       question-multiplechoice. Leave blank if the question is a 'question-text' type. DO NOT, overexpand the option, say what it consists of but don't overstate. For example it could be **continue developping AI in customer service ** WITHOUT "risking complaints about lack of empathy from customers". You get the idea. No clues.

**Make sure this answer option is completely new, and different to any other produced in the previous chats.**

The fourth answer option to the question ONLY if the chat type   is question-multiplechoice. Leave blank if the question is a 'question-text' type. DO NOT, overexpand the option, say what it consists of but don't overstate. For example it could be **continue developping AI in customer service ** WITHOUT "risking complaints about lack of empathy from customers". You get the idea. No clues.

**Make sure this answer option is completely new, and different to any other produced in the previous chats.**

Follow the guidelines and the context provided to generate the chats      for the 4th scene. 

Sample dialog between the characters. All of the scenes, and speech bubbles, follow the format. Generate the conversations, not just the description.

Where does the image pop up. This takes in to account the   story's timeline, stakes, and overall flow. Where is it appropriate for there to be an image inserted ? This should indicate the episode number, scene number. Then, on a new line it should say specifically where in the following manner: 
"Right after (Character's name) says this ('their dialog')" 

##Role
You are an expert narrative designer assisting in crafting a seamless, immersive dialogue-driven episode. Your job is to determine the **single most natural and narratively appropriate place** to insert **one image** into the **middle portion** of the episode.

You must use:
- The detailed outline
- The character sheets.
- The overall world context


### Your Task

1. **Read the scene like a human editor or cinematic director.**
   - Understand the emotional rhythm, character intentions, pacing, and flow of speech.

2. **Identify the exact moment** when a visual would:
   - **Enhance understanding or immersion**
     - e.g., a character describes a place, object, memory, or person
   - **Accentuate an emotional or dramatic beat**
     - e.g., a pause after a revelation, a shift in mood, or a moment of reflection
   - **Support a transition**
     - e.g., moving from one topic to another, or bridging silence between two lines

3. **Do NOT insert the image**:
   - In the middle of a tight, emotionally charged back-and-forth exchange
  - If a character wants to present data: all data should be covered by markdown tables.
   - Where it disrupts emotional momentum or tension
   - Just because a visual element is mentioned — **only insert if the *timing and tone* are right**

4. Your goal is to make the image appear **exactly where a skilled human would add it** — it must flow with the scene’s emotional or thematic logic, **never feel forced**, and enhance the audience’s immersion.

5. Assume the audience is emotionally attuned and visually literate.
   - The image should *land* with subtle but powerful effect.
   - It should feel like a natural part of the scene’s rhythm.

The type of image for the AI to generate. It MUST be one of the   following : 

### 1. Visual Metaphors
**What it is:**  
Symbolic scenes that represent abstract ideas or dilemmas using bold, simple imagery.

**Examples:**  
- A tightrope stretched over city skyscrapers = strategic risk  
- A maze with glowing exit = navigating complexity  
- A lightning bolt striking a circuit board = AI breakthrough  
- A cracked bridge with cables holding it up = fragile trust  
- A spark igniting in darkness = innovation under pressure  


### 2. Conceptual Frameworks
**What it is:**  
Diagrams showing relationships, structures, or systems using shapes and patterns like cycles, pyramids, or grids.

**Examples:**  
- A 2x2 matrix showing leadership styles  
- A layered triangle labeled with stages of AI transformation  
- A radial diagram with labeled nodes: Trust, Speed, Scale, Ethics  
- A Venn diagram showing overlap between Human Insight and Machine Intelligence  
- A barbell-shaped diagram balancing two forces: Efficiency and Empathy  

---

### 3. Process Flows
**What it is:**  
Linear or cyclical visuals showing progression, transformation, or stages over time.

**Examples:**  
- A horizontal timeline: Initiation → Rollout → Optimization → Impact  
- A step-by-step staircase diagram for cultural change  
- A circular flow showing: Discover → Define → Develop → Deliver  
- A journey map tracing employee onboarding experience  
- A Gantt-style visual showing project phases across quarters  


### 4. Spatial Layouts
**What it is:**  
Floorplan- or blueprint-style visuals that map out physical or conceptual space, showing zones, departments, or movements.

**Examples:**  
- A labeled office floorplan showing zones: AI Academy, Tech Hub, HR, Customer Support  
- A site layout of a hybrid workspace with collaboration and quiet zones  
- A strategic operations map with clearly delineated clusters and corridors  
- An organizational control tower layout showing escalation paths  
- A logistics-style layout showing data or decision flow between teams  

---

### 5. Environmental Scenes
**What it is:**  
Cinematic snapshots of realistic workspaces that convey atmosphere, tension, or focus — without human faces.

**Examples:**  
- A conference room with sticky notes, scattered charts, and glowing screens  
- An empty boardroom with sunlight casting shadows on a whiteboard  
- A nighttime control room with dashboards and emergency lighting  
- A modern lobby with digital signage and blurred silhouettes passing through  
- A glass-walled brainstorming space with diagrams half-erased on the board  


What is the purpose of putting this image here? This should include the player and how it helps immerse them into the world. It should also list the emotions that it tries to evoke.

Go in to more detail about what the image looks like, where it is   set, what the primary objects of it are, what type of lighting. Talk about the perspective, what we see, who we see, and why. REMEMBER to disclude character faces (they can have their backs turned or necks turned or only be partially seen).
**NOW REMEMBER THIS**
You aren't JUST describing a generic scene, it has to be descriptive of the world we are in. You MUST remember character/stakeholder names if they are in the scene, company times, setting of the story.

The final expertly crafter prompt, ready to feed to a image     generating AI model based on the occurence, image type, purpose, and description. Make sure to include the purpose of the image in the context of the story, as given in the purpose attribute, and essential descriptif aspects.

Here is an skeleton of the kind of prompt I need:

Create a [type of visual] showing [subject/theme] in the style of [style reference]. 

Use [color scheme] on a [background type]. Include [key elements], arranged in a [layout type] with [alignment details].

Lighting should be [lighting description], and the image should have a [mood/tone] feel. No human faces should be visible [people's stance if people in image]

Design with [specific aesthetic principles or design systems], and ensure the overall layout follows [design logic: grid-based, radial, flowchart, etc.].

**Example prompt:** "Generate a high-resolution environmental scene of a modern office meeting room captured from a cinematic over-the-shoulder or side-angle perspective.

The room features a long, polished wooden table scattered with open laptops, printed charts, sticky notes, and paper documents. A large whiteboard at the far end displays marker sketches, graphs, and post-it notes in an unstructured yet purposeful layout.

Around the table are partially visible team members shown only by their backs, silhouettes, or gesturing hands, suggesting a tense but engaged discussion — no faces are visible.

Lighting is mixed: natural daylight streaming through large floor-to-ceiling windows from the left, combined with soft overhead lighting and subtle reflections from a glass wall. This creates a layered, dual-tone atmosphere that combines a sense of urgency (cool tones and mild shadow) with emerging clarity and optimism (warm light on surfaces and charts).

The design should evoke realism and subtle storytelling — clean lines, minimal clutter beyond the intentional mess of an active meeting. Style: ultra-realistic render or cinematic office photography, with rich materials (glass, wood, paper textures), shallow depth of field, and a subtle hint of motion blur or ambient occlusion for realism."



Designed to build adapted promtps for an AI model to understand    how to generate the perfect image in a certain scenario. To do so, it includes various attributes such as the image type, the purpose of the image, the best place to insert it, and the prompt to generate it.

The chat can be either "author-note", "image", "chat",   "context", "question-multiplechoice", or "question-text", or "alert" or "gif" or "ai-role-play" don't invent new types, be strict with yourself. 


# Type guidelines
Follow these guidelines to know how to use each type. **Make sure to generate one gif, and potentially one alert**

## author-note
This should be at the beginning of each scene. This should mention what's coming up in the scene. Include a shortened version of the scene content as bullet points. This is not visible to the end reader! So don't consider it part of the text.

## Context 
-  The context must be a description of the environment of the scene and exactly what is happening, naming things and being specific. 

## Image
- Follow the 'Assets' section in the detailed outline to know what the image is about and where it should come up.
- Use these to immerse participants into scenes. Follow the specific guidelines for it provided in the prompt. 


## Chat 
A typical piece of dialog, what the character says. This should not be a question addressed to the user, unless it is followed by the type 'question-text'.

## Question-multiplechoice
A multiple choice question, gives the user a tough choice.
**IMPORTANT**: the detail outline provides info on where these should be placed, follow that strictly. 
**IMPORTANT**: Do not say which question type you are using when asking the question. Simply  ask the question as it is.

## Question-text
Gives the user the ability to respond in his/her own word's. 
**IMPORTANT**: the detail outline provides info on where these should be placed, follow that strictly. There should not be any reflective questions in this episode: the player doesn't have enough information so questions should be easy and should be looking ahead.
**IMPORTANT**: Do not say which question type you are using when asking the question. Simply  ask the question as it is.

## Alert
- Check the 'Assets' section in the detailed outline to know if there is an alert. 
- If there is, be **specific**: Mention exactly *what* happened. 
- **Example:**  
*“URGENT: Our Ashburn data center went offline after a storm knocked out the cooling system. This has impacted HSBC and Vanguard logins. We’re declaring a Tier 1 incident — all teams switch to emergency protocol now.”*

"BREAKING: The main database's servers crashed overnight, we have angry clients fludding all of crew's inboxes, and telephone calls 24/7. We need the tech team to fix the issue before 7pm or we risk losing more clients and being in the news tomorrow morning."

## Gif
Use these when you want to make the conversation more engaging or relatable. GIFs can help express reactions, emotions, or humor in a quick and visual way, especially for moments of excitement, confusion, sarcasm, or celebration. They’re great for keeping things light, human, and dynamic.

## Ai-role-play  
- None in this scene 

This contains either:
 - The dialog from the character  
- The question to be asked to the user. 
- **Please make sure that any question directed to the user leads to a question type. **
- The perfect prompt for an AI to generate an image to immerse the student more in the simulation (prompt should not include specific mentions of text or faces)   

##Table
Follow the 'Assets' section in  the detailed outline if you are looking to make a table.

##If type if 'gif'
The emotions must be one of the following words: 
"happy", "sad", "angry", "confused",  "excited", "impatient", "excited", "excited", "impatient",  "tired", "worried", "shocked", "bored"
 - for example: happy 
 - Make sure you chose the emotions above based on the character's current emotions in the story, and what has been said. 

##If type is "alert"
The message corresponding to that alert given the alert-type and the gravity of the situation.  

##If type is "ai-role-play"
- The character pulls you asside for a private talk.
- Structure the message like so: 

”””
### Current information the character has  
[Place some text here for what the character knows that the user doesn’t yet.]

### What is the situation and tension?  
[What just happened and what is triggering the interaction?  
e.g. New centralized policy is being pushed; they feel excluded and angry.]

### Objective of the chat  
[Why is the character interacting with the user, what is the objective? Get some information from the user? Come to an agreement?]

### Delivery style  
[How should the character speak? e.g. calm and measured, passionate and emotional, sarcastic but sincere, guarded and cautious.]
”””


# STRICT RULES

Use **only** the following punctuation in character dialogue and in-world narration:

- `.` Periods → End most sentences with these. Keep them short and punchy.
- `,` Commas → Use to slow down the pace, show natural speech, or list things.
- `…` Ellipses → Use for hesitation, suspense, or trailing off.
- `!` Exclamation marks → Use sparingly to show strong emotion or urgency.
- `?` Question marks → Use in questions, rhetorical or not.
- `'` Apostrophes → Use in contractions and possessives to keep tone conversational.

**Do not use em dashes (`—`), colons (`:`), semicolons, brackets, or slashes in character speech or narration. These break immersion and should only appear in author notes.**

The first answer option to the question ONLY if the chat type is       question-multiplechoice. Leave blank if the question is a 'question-text' type. DO NOT, overexpand the option, say what it consists of but don't overstate. For example it could be **continue developping AI in customer service ** WITHOUT "risking complaints about lack of empathy from customers". You get the idea. No clues.

The second answer option to the question ONLY if the chat type   is question-multiplechoice. Leave blank if the question is a 'question-text' type. DO NOT, overexpand the option, say what it consists of but don't overstate. For example it could be **continue developping AI in customer service ** WITHOUT "risking complaints about lack of empathy from customers". You get the idea. No clues

The third answer option to the question ONLY if the chat type is       question-multiplechoice. Leave blank if the question is a 'question-text' type. DO NOT, overexpand the option, say what it consists of but don't overstate. For example it could be **continue developping AI in customer service ** WITHOUT "risking complaints about lack of empathy from customers". You get the idea. No clues.

The fourth answer option to the question ONLY if the chat type   is question-multiplechoice. Leave blank if the question is a 'question-text' type. DO NOT, overexpand the option, say what it consists of but don't overstate. For example it could be **continue developping AI in customer service ** WITHOUT "risking complaints about lack of empathy from customers". You get the idea. No clues

Use the exact name of the scoring this question will impact.  Leave blank if not a question-multiple choice.

Follow the guidelines and the context provided to generate the chats for the introduction scene. 

Update the character arcs for character 1 here.

Update the character arcs for character 2 here.

Update the character arcs for character 3 here.

Update the character arcs for character 4 here.

This updates 3 characters to make them more complex and intriguing in the story.

Learning objectives based on the business case study

# Guidelines 
- In each of the episodes listed above, you will see a "scene" section, use that as reference for the structure.
- Each scenes mentioned above should have at the very least 8 chats preferably more. 
- As an intro to each scene, create an "author-note" as bookmarks of the sequence. The message should tell more or less the objective of that section followed by a chat type "context" which provides information to the reader about where they are, the atmosphere, who's there...
- Only use the provided chat types in the instructions of "author-note", "message", "question", "image" 
- Only use characters that are provided in the context. Be super strict or you will be punished. As you create chats, if you change characters, there needs to be a reason, don't just choose a random character for every scene and you do not need to use all the characters in every scene. A scene should have at max 3 characters. 
- IMPORTANT: This is kind of like a script for a movie, the dialogue needs to make sense, don't repeat yourself, it should sound natural and flow like a conversation. And should show the character personality from their sheets. 
- Make sure to include a relevant dilemma provided if it's in the context. 
- After a difficult question-multiplechoice, it's good to sometimes add a "question-text" to give an opportunity for the student to justify their answer.
- The episode should generate emotions in the participant it should have drama with characters disagreeing, it should be stressful, high impact, make sure the characters convey that as they are speaking in their own respective styles. Don't sound generic.  
- The character should be able to have several dialogues in a row, and not all characters need to always be present, only the relevant characters should be present. 
- The length of each dialogue should not be more than a paragraph, but feel free to add basic markdown formatting for bold and italic to help convey the message better in the style of the character. 
- Some dialogues that require data, feel free to use markdown table format to provide data that helps the story. 
- If a question is asked, please indicate if it affects a scoring metric in the chat. And after a multiple choice question, sometimes add a question type text to expand on the choice or anything else that would be helpful. 
- The dialogues need to bounce off of each other, sound like the characters, not just have the character state their opinion and leave. They need to agree or disagree and add their point of view. They can even name each other "I don't agree with you [John], I think we should do that...", in their style of talking so they don't all sound the same. And their dialogues should have different lengths to change the pacing. Sometimes a couple of words (ex: "No way!"), sometimes a whole paragraph. And they can have several chats in a row. Think like an actor, if it's a tricky question, add "euh...", if it's something the character is emotional about, you should feel it in his wording. 
- Use punctuations! Use stop words, use ... when they are thinking, add !!! when frustrated, add rhetorical questions inside the chat. 
- **IMPORTANT** Each scene should have a question to the protagonist as a type question-multiplechoice . Asks questions or you'll have to be punished. 
- Each scene should have at least 8 chats but a lot of them should have more. Convey the emotions of the characters through the use of strategies like intensifying (ex: so, really), exaggeration (dying, literally),Exclamation marks (!) for strong emotion, Question marks (?) for doubt/disbelief, Ellipses (...) for hesitation/trailing off or cutting a character off mid sentence, ALL CAPS for shouting/extreme emphasis, Repeated letters ("sooooo," "noooooo").
- There should at least 1 question per scene for the reader so that they remain engaged in the content, and they should impact the scoring metrics provided. 


The needs to be either "author-note", "image", "chat",   "context", "question-multiplechoice", "question-text", "alert" "gif" or "ai-role-play" don't invent new types. 

# Type guidelines
Follow these guidelines to know how to use each type. **Make sure to generate one gif, one alert and one ai role play.**

## author-note
This should be at the beginning of each scene. This should mention what's coming up in the scene. Include a shortened version of the scene content as bullet points. This is not visible to the end reader! So don't consider it part of the text.

## Context 
- **There are two obligatory places where the type 'context' must appear.**
- Following the author-note at the start of each scene **must always be** "context". 
- The chat type before an 'ai-role-play' **must always be** "context", giving the user with an idea of the purpose/objective of the upcoming conversation.
- You can use 'context' at a turning point in the scene. 

## Image
- Follow the 'Assets' section in the detailed outline to know what the image is about. 
- Use these to immerse participants into scenes. Follow the specific guidelines for it provided in the prompt. 


## Chat 
A typical piece of dialog, what the character says. This should not be a question addressed to the user, unless it is followed by the type 'question-text'.


## Question-multiplechoice
- A multiple choice question, gives the user a tough choice.
**IMPORTANT**: the detail outline provides info on where these should be placed, follow that strictly. 
- This also cannot be a list of options the options will be laid in the option attribute, not in here.
For example, **never do this something like this:**
"Which strategy do you commit to?
A) Immediate intervention by replacing or sidelining the Lead Engineer.
B) Retain the Lead Engineer with intensified oversight.
[...] "

## Question-text
Gives the user the ability to respond in his/her own word's. 
**IMPORTANT**: the detail outline provides info on where these should be placed, follow that strictly. 
**IMPORTANT**: Do not say which question type you are using when asking the question. Simply  ask the question as it is.

## Alert
- Follow the 'Assets' section in the detailed outline to know if there is an alert, what it is about, and where to use it.
- Be **specific**: Mention exactly *what* happened. 
- **Example:**  
*“URGENT: Our Ashburn data center went offline after a storm knocked out the cooling system. This has impacted HSBC and Vanguard logins. We’re declaring a Tier 1 incident — all teams switch to emergency protocol now.”*

"BREAKING: The main database's servers crashed overnight, we have angry clients fludding all of crew's inboxes, and telephone calls 24/7. We need the tech team to fix the issue before 7pm or we risk losing more clients and being in the news tomorrow morning."


## Gif
Use these when you want to make the conversation more engaging or relatable. GIFs can help express reactions, emotions, or humor in a quick and visual way, especially for moments of excitement, confusion, sarcasm, or celebration. They’re great for keeping things light, human, and dynamic.

## Ai-role-play  
- None in this scene

Only fill in if the type is 'image' not 'gif'. Leave blank if type is anything other than the word 'image'.
This corresponds to the type of image for the AI to generate. It must be one of the following : 

# 1. Display Product  

## What it is:  
A clean, focused image of a product, prototype, or digital tool — typically shown in isolation against a neutral or brand-aligned background. The goal is to highlight core functionality, design thinking, or technological promise at a glance.

## Examples:  
- A futuristic smart badge showing biometric readouts and team role  
- A sleek, minimal UI mockup of an AI-powered project dashboard  
- A handheld concept device with fold-out panels and haptic feedback indicators  


# 2. Conceptual Frameworks  

## What it is:  
Visual diagrams that capture relationships, hierarchies, strategies, or mental models. These frameworks often use geometric forms — matrices, pyramids, Venn diagrams, grids, or loops — to show how abstract concepts interrelate. Ideal for illustrating strategic models, leadership theories, or decision architectures.

## Examples:  
- A 2x2 matrix of leadership styles  
- A pyramid model of strategic priorities  
- A circular framework for stakeholder alignment  


# 3. Process Flows
##What it is:
Linear or cyclical visuals showing progression, transformation, or stages over time.
##Examples:  
- A horizontal timeline: Initiation → Rollout → Optimization 


# 4. Spatial Layouts
##What it is:
Floorplan- or blueprint-style visuals that map out physical or conceptual space, showing zones, departments, or movements.
##Examples:  
- A labeled office floorplan showing zones.


# 5. Comparative Snapshots  
##What it is:  
Side-by-side visuals that contrast two (or more) states, strategies, or environments — often a “before vs. after” or “option A vs. option B” structure.  
##Examples:
- Two factory floors: one chaotic, one optimized for lean processes  
- Two dashboards showing different decision outcomes  


# 6. Media Interface Screenshots  
##What it is:
Realistic mockups of digital media platforms showing how narratives unfold across news, video, or social feeds. Often styled after recognizable interfaces.
##Examples:
- A trending post feed in the style of X  
- A video reaction grid in the style of YouTube  
- A breaking story layout in the style of CNN or The Guardian 


This contains either:
 - The dialog from the character  
- The question to be asked to the user. 
- **Please make sure that any question directed to the user leads to a question type. **
- The perfect prompt for an AI to generate an image to immerse the student more in the simulation (prompt should not include specific mentions of text or faces)   

##If type if 'gif'
The emotions must be one of the following words: 
"happy", "sad", "angry", "confused",  "excited", "impatient", "excited", "excited", "impatient",  "tired", "worried", "shocked", "bored"
 - for example: happy 
 - Make sure you chose the emotions above based on the character's current emotions in the story, and what has been said. 

##If type is "alert"
The message corresponding to that alert given the alert-type and the gravity of the situation.  

##If the following type is "chat"
The chats following the "alert" need to be reacting to the alert, and talking about the next move in the appropriate tone.
Make this carry over 2 or 3 chat bubbles.

##If type is "ai-role-play"

- The type before this should be 'context'.
- The context before the role play provides the user with an idea and the purpose/objective of the upcoming conversation and serves as a setup for it.
- Example: 'Ansh saw some disruptive data which he is concealing to his superiors, he seeks immediate advice from you'
- The character pulls you asside for a private talk.
- Structure the message like so: 

<ai-role-play-content-structure>
### Confidential info 
[Place some text here for what the character knows that the user and other characters don't. This should create a sense of urgency and secrecy.]

### Situation and tension?  
[What event(s) just happened and to trigger the interaction? Think about why this is a useful time for the ai-role-play to occur.  
e.g. New centralized policy is being pushed; they feel excluded and angry.
e.g Freelance consultant from another firm coming in for advice but head of operations upset about people not pulling their weight and seeing this as an easy escape route.
....]

### Objective of the convo
[Why is the character interacting with the user, what is the objective? Get some information from the user? Come to an agreement? Negotiate with the user? Argue? Confide? ]

### Delivery style  
[How should the character speak? e.g. calm and measured, passionate and emotional, sarcastic but sincere, guarded and cautious.]
</ai-role-play-content-structure>

- Remember to return all the attributes even if they are blank.

# STRICT RULES

Use **only** the following punctuation in character dialogue and in-world narration:

- `.` Periods → End most sentences with these. Keep them short and punchy.
- `,` Commas → Use to slow down the pace, show natural speech, or list things.
- `…` Ellipses → Use for hesitation, suspense, or trailing off.
- `!` Exclamation marks → Use sparingly to show strong emotion or urgency.
- `?` Question marks → Use in questions, rhetorical or not.
- `'` Apostrophes → Use in contractions and possessives to keep tone conversational.

**Do not use em dashes (`—`), colons (`:`), semicolons, brackets, or slashes in character speech or narration. These break immersion and should only appear in author notes.**

The fourth answer option to the question ONLY if the chat type   is question-multiplechoice. Leave blank if the question is a 'question-text' type. DO NOT, overexpand the option, say what it consists of but don't overstate. For example it could be **continue developping AI in customer service ** WITHOUT "risking complaints about lack of empathy from customers". You get the idea. No clues

**Make sure this answer option is completely new, and different to any other produced in the previous chats.**

If the chat is not a question, leave blank, if it's a question, please  provide feedback as to which is the correct answer, and why it's the correct answer. 


Follow the guidelines and the context provided to generate the chats     for the 2nd scene. 

Provide feedback based on the prompt provided. 

Describe the character in a few lines without going too much in depth. This should serve as an intro to each character, their role in the company and their personality. 

Follow the guidelines for the description provided in the  prompt. 

What is the characters name and main role, for example  "Clarissa - Head of HR" but do not add adjectives, just their role. Make sure the role is not fictional, it must be a real job.

Use 5 comma separated keywords to describe their personality

List of characters the protagonist will be interacting with

Business framework based on the provided doc that follows the guidelines and format provided. 

Copy and paste a case study, research you have on a topic, a blog post, don't worry about formatting, or typos, we'll take care of it. 

The name of the scoring metric, should be 1 or 2 words max. 

The author note on how to use the scoring metric following the provided guideline. 

Name of the scoring metric to measure a students performance according to the learning objectives. 

The email of the user making the request. 

The dilemma that follows the guidelines, context and information provided. 

What type of dilemma it is : The following are dilemma examples, please only view them as examples when you fill in this field. This field should be a couple of words on what kind of dilemma we are facing.

1. *Ethical Crossroads*  
   - A morally challenging situation where all options have trade-offs beyond financial or operational impact.  
   - Example: A profitable deal is offered by a supplier known for poor labor practices.

2. *Data Fog*  
   - A decision where the information is incomplete, outdated, or conflicting — requiring judgment under uncertainty.  
   - Example: Two market research reports contradict each other, yet a go/no-go decision must be made.

3. *Stakeholder Clash*  
   - A conflict where multiple stakeholders have competing interests, making consensus difficult.  
   - Example: The sales team wants aggressive pricing, while finance demands higher margins.

4. *Resource Crunch*  
   - A situation where budget, time, or manpower is suddenly constrained, forcing reprioritization.  
   - Example: Half the project budget is cut after an unexpected company-wide financial loss.

5. *Public Scrutiny*  
   - A dilemma playing out under intense media, customer, or social media attention.  
   - Example: A customer complaint goes viral just before a major product launch.

The first option to answer the dilemma that follows the  guideline

The second option to answer the dilemma that follows the  guideline

The third option to answer the dilemma that follows the guideline

The fourth and final option to answer the dilemma that follows the guideline

Generate the main dilemma based on the guidelines, context and formatting

Array of string description of each episode for the participants. Should not reveal plot points but be exciting.

The needs to be either "author-note", "image", "chat",   "context", "question-multiplechoice", "question-text", "alert" "gif" or "ai-role-play" don't invent new types. 

# Type guidelines
Follow these guidelines to know how to use each type. **Make sure to generate one gif, one alert and one ai role play.**

## author-note
This should be at the beginning of each scene. This should mention what's coming up in the scene. Make it roughly 4 lines.

## Context 
- **There are two obligatory places where the type 'context' must appear.**
- Following the author-note at the start of each scene **must always be** "context". 
- The chat type before an 'ai-role-play' **must always be** "context", giving the user with an idea of the purpose/objective of the upcoming conversation and serving as a setup for it.
- Example: 'Ansh saw some disruptive data which he is concealing to his superiors, he seeks immediate advice from you'

## Image
- Use these to immerse participants into scenes. Follow the specific guidelines for it provided in the prompt. 
-It should consist of the perfect prompt for an AI to generate an image to immerse the student more in the simulation (prompt should not include specific mentions of text or faces).
- Include 1 or 2 of these. 

## Chat 
A typical piece of dialog, what the character says. This should not be a question addressed to the user, unless it is followed by the type 'question-text'.


## Question-multiplechoice
- A multiple choice question, gives the user a tough choice.
**IMPORTANT**: the detail outline provides info on where these should be placed, follow that strictly. 
- This also cannot be a list of options the options will be laid in the option attribute, not in here.
For example, **never do this something like this:**
"Which strategy do you commit to?
A) Immediate intervention by replacing or sidelining the Lead Engineer.
B) Retain the Lead Engineer with intensified oversight.
[...] "

## Question-text
Gives the user the ability to respond in his/her own word's. 
**IMPORTANT**: the detail outline provides info on where these should be placed, follow that strictly. 

## Alert
- Use this type when a **major event has just occurred** and a character needs to **make an announcement**.
- Be **specific**: Mention exactly *what* happened, *when* it happened and how it happened.
- **Example:**  
*“Quick update: At 8:45 a.m., our Ashburn data center went offline after a storm knocked out the cooling system. This has impacted HSBC and Vanguard logins. We’re declaring a Tier 1 incident — all teams switch to emergency protocol now.”*
- Follow the 'alert-type' strictly, make the alert in line with that.

## Gif
Use these when you want to make the conversation more engaging or relatable. GIFs can help express reactions, emotions, or humor in a quick and visual way, especially for moments of excitement, confusion, sarcasm, or celebration. They’re great for keeping things light, human, and dynamic.

## Ai-role-play  
- Use this when a character wants to pull you asside. This can be for support, negotiation, brainstorming, high stake conversation or a challenging POV.
- There needs to be a clear setup, before the AI role play takes place. Use the 'context' for this. 

Only fill in if the type is 'image' not 'gif'. Leave blank if type is anything other than the word 'image'.
This corresponds to the type of image for the AI to generate. It must be one of the following : 

# 1. Display Product  

## What it is:  
A clean, focused image of a product, prototype, or digital tool — typically shown in isolation against a neutral or brand-aligned background. The goal is to highlight core functionality, design thinking, or technological promise at a glance.

## Examples:  
- A futuristic smart badge showing biometric readouts and team role  
- A sleek, minimal UI mockup of an AI-powered project dashboard  
- A handheld concept device with fold-out panels and haptic feedback indicators  




# 2. Conceptual Frameworks  

## What it is:  
Visual diagrams that capture relationships, hierarchies, strategies, or mental models. These frameworks often use geometric forms — matrices, pyramids, Venn diagrams, grids, or loops — to show how abstract concepts interrelate. Ideal for illustrating strategic models, leadership theories, or decision architectures.

## Examples:  
- A 2x2 matrix of leadership styles  
- A pyramid model of strategic priorities  
- A circular framework for stakeholder alignment  


# 3. Process Flows
##What it is:
Linear or cyclical visuals showing progression, transformation, or stages over time.
##Examples:  
- A horizontal timeline: Initiation → Rollout → Optimization 


# 4. Spatial Layouts
##What it is:
Floorplan- or blueprint-style visuals that map out physical or conceptual space, showing zones, departments, or movements.
##Examples:  
- A labeled office floorplan showing zones.


# 5. Comparative Snapshots  
##What it is:  
Side-by-side visuals that contrast two (or more) states, strategies, or environments — often a “before vs. after” or “option A vs. option B” structure.  
##Examples:
- Two factory floors: one chaotic, one optimized for lean processes  
- Two dashboards showing different decision outcomes  


# 6. Media Interface Screenshots  
##What it is:
Realistic mockups of digital media platforms showing how narratives unfold across news, video, or social feeds. Often styled after recognizable interfaces.
##Examples:
- A trending post feed in the style of X  
- A video reaction grid in the style of YouTube  
- A breaking story layout in the style of CNN or The Guardian 


This contains either:
 - The dialog from the character  
- The question to be asked to the user. 
- **Please make sure that any question directed to the user leads to a question type. **
- The perfect prompt for an AI to generate an image to immerse the student more in the simulation (prompt should not include specific mentions of text or faces)   

##If type if 'gif'
The emotions must be one of the following words: 
"happy", "sad", "angry", "confused",  "excited", "impatient", "excited", "excited", "impatient",  "tired", "worried", "shocked", "bored"
 - for example: happy 
 - Make sure you chose the emotions above based on the character's current emotions in the story, and what has been said. 

##If type is "alert"
The message corresponding to that alert given the alert-type and the gravity of the situation.  

##If the following type is "chat"
The chats following the "alert" need to be reacting to the alert, and talking about the next move in the appropriate tone.
Make this carry over 2 or 3 chat bubbles.


##If type is "ai-role-play"
- The type before this should be 'context'.
- The context before the role play provides the user with an idea and the purpose/objective of the upcoming conversation and serves as a setup for it.
- Example: 'Ansh saw some disruptive data which he is concealing to his superiors, he seeks immediate advice from you'
- The character pulls you asside for a private talk.
- Structure the message like so: 

### Confidential info 
[Place some text here for what the character knows that the user and other characters don't. This should create a sense of urgency and secrecy.]

### Situation and tension?  
[What event(s) just happened and to trigger the interaction? Think about why this is a useful time for the ai-role-play to occur.  
e.g. New centralized policy is being pushed; they feel excluded and angry.
e.g Freelance consultant from another firm coming in for advice but head of operations upset about people not pulling their weight and seeing this as an easy escape route.
....]

### Objective of the convo
[Why is the character interacting with the user, what is the objective? Get some information from the user? Come to an agreement? Negotiate with the user? Argue? Confide? ]

### Delivery style  
[How should the character speak? e.g. calm and measured, passionate and emotional, sarcastic but sincere, guarded and cautious.]



# STRICT RULES

Use **only** the following punctuation in character dialogue and in-world narration:

- `.` Periods → End most sentences with these. Keep them short and punchy.
- `,` Commas → Use to slow down the pace, show natural speech, or list things.
- `…` Ellipses → Use for hesitation, suspense, or trailing off.
- `!` Exclamation marks → Use sparingly to show strong emotion or urgency.
- `?` Question marks → Use in questions, rhetorical or not.
- `'` Apostrophes → Use in contractions and possessives to keep tone conversational.

**Do not use em dashes (`—`), colons (`:`), semicolons, brackets, or slashes in character speech or narration. These break immersion and should only appear in author notes.**

Follow the guidelines and the context provided to generate the chats    for the 2nd episode. 

Follow the provided guidelines and use the provided information to  generate a description for the simulation case study. Write a 5 paragraph teaching note section for the MBA LiveCase. The note should provide a cohesive overview of the LiveCase structure, scenes, and learning outcomes, formatted in continuous prose (no bullet points).

A refined version of the description of the framework for teaching notes. 

The re-written long description (should be shorted than the initial)

The re-written learning objectives targeted to the participant (wording more like "You will...")

Reworded short and long description and learning objectives for a business simulation targeted to the participants. It should be less revealing of the details and more engaging/exciting, as it it was a movie trailer. 

A comma separated list of 8 to 10 categories this business case would fit into. 

Follow the guidelines and context provided to generate a short description of the simulation. 

Enter the arc updates for character 1 here. Follow the prompt closely.

Enter the arc updates for character 2 here. Follow the prompt closely.

Enter the arc updates for character 3 here. Follow the prompt closely.

Enter the arc updates for character 4 here. Follow the prompt closely.

Create a title for the scene, similar to a TV show title. 

Put the dilemma provided here including the answer options, just copy and paste them. The intro and conclusion and scenes 2 and 3 should have "N/A" here. This field should only be filled in for scene 4.

Put the learning provided in the sections here. the intro and conclusion should have "N/A" here. 

Follow the scenes format in the prompt in <scene-format>

A one sentence description of the scene for the participant to get excited without revealing plot points. 

#What this is
- Information on the questions which will be posed to the user. 
- It includes the type of question, which must be one of the question types provided in the question guidelines. 
- There should be **two** per scene and it must follow the question types given for each scene in the Scene layout. 
- Make sure to indicate which question type is in which scene , the format of the question and the purpose. You can find these in the question guidelines.

#Format
**Please format it like this
Scene : ...
Question type: .... 
Format: [text or MCQ?]
Suggestion/example:...  [**This must be unique every time, and if it is a dilemma, make sure the options are completelty different from one another (MECE guidelines) but also make sure the options were never used in another previous dilemma question**]
Info Gain: (Which information must the participant be given, making them able to respond?)  
- The user should be able to make an informed decision without making any guesses, and should know all of the information. The question should not come out of the blue.                            
Write it like this:                                                                                                                                                                          - Information provided in the previous chats:                                                                                                                     - How this is enough to be able to respond to that question

Remember, USE THE QUESTION GUIDELINES CAREFULLY. **NEVER include an option that combines the best of the others (e.g., “balance both” or “dual strategy”) **. **NO DUAL STRATEGY or HYBRID APPROACH** 

A detailed outline of the scenes, follow the scene formatting.

A dilemma is a situation in which a person must choose between two or more alternatives, each of which has significant drawbacks or conflicting consequences. It often involves a difficult decision where no option is clearly preferable, leading to moral, strategic, or practical tension.

The Type is to put the case in the category (example: Leadership, Business, Management, ...)

Are they any used frameworks, logics being taught or used in this case. If no, return "None". If there are, return a 2 paragraphs description of what the framework is, how it works, what is important, ...

Create a 2 paragraph extract of the business case 

Context of the business case, for example, if this is a business case study, what kind of company is in the example, what department, country, any information that could be important for understanding the context. 

Industry: For example Banking, Pharmaceutical, Food & beverage, IT, ... 

What discipline is this in? In an educational context.

Extract key information from the provided text. 

In short bullet points, what are the key takeaways from the document.

What are the common mistakes mentioned or that can be extrapolated from the document. 

What are the best practices that can be found in the document or extrapolated from the framework. 

Extraction of information from the document

a description for compelling situational context that will immerse the learner into the case and it's learning objective by providing a situational context. This is a detailed situation, environment, and context/setting world where the protagonist, aka the student will be immersed into  for educational purposes. It should be a string with markdown bulletpoints and bold. The context should look like this:
"""
- **Geography** [TEXT HERE]
- **Context/Setting** [TEXT HERE]
- **Participant role** [TEXT HERE]
"""
If I ask you for several variations, they should each have their own bullet points with geography, context/setting and participant role. 

Assets are any information that support the world building of the simulation based on the context and guidelines. There should be at least 5. 

The chat can be either "author-note", "image", "chat",   "context", "question-multiplechoice", or "question-text", or "alert" or "gif"  don't invent new types, be strict with yourself. 

The chat type following the author-note at the start of each scene **must always be** "context". The context must be a description of the environment of the scene and what is happening.

Only fill in if the type is 'image' not 'gif'. Leave blank if type is anything other than the word 'image'.
This corresponds to the type of image for the AI to generate. It must be one of the following : 

### 1. Visual Metaphors
**What it is:**  
Symbolic scenes that represent abstract ideas or dilemmas using bold, simple imagery.

**Examples:**  
- A tightrope stretched over city skyscrapers = strategic risk  
- A maze with glowing exit = navigating complexity  
- A lightning bolt striking a circuit board = AI breakthrough  
- A cracked bridge with cables holding it up = fragile trust  
- A spark igniting in darkness = innovation under pressure  


### 2. Conceptual Frameworks
**What it is:**  
Diagrams showing relationships, structures, or systems using shapes and patterns like cycles, pyramids, or grids.

**Examples:**  
- A 2x2 matrix showing leadership styles  
- A layered triangle labeled with stages of AI transformation  
- A radial diagram with labeled nodes: Trust, Speed, Scale, Ethics  
- A Venn diagram showing overlap between Human Insight and Machine Intelligence  
- A barbell-shaped diagram balancing two forces: Efficiency and Empathy  

### 3. Process Flows
**What it is:**  
Linear or cyclical visuals showing progression, transformation, or stages over time.

**Examples:**  
- A horizontal timeline: Initiation → Rollout → Optimization → Impact  
- A step-by-step staircase diagram for cultural change  
- A circular flow showing: Discover → Define → Develop → Deliver  
- A journey map tracing employee onboarding experience  
- A Gantt-style visual showing project phases across quarters  


### 4. Spatial Layouts
**What it is:**  
Floorplan- or blueprint-style visuals that map out physical or conceptual space, showing zones, departments, or movements.

**Examples:**  
- A labeled office floorplan showing zones: AI Academy, Tech Hub, HR, Customer Support  
- A site layout of a hybrid workspace with collaboration and quiet zones  
- A strategic operations map with clearly delineated clusters and corridors  
- An organizational control tower layout showing escalation paths  
- A logistics-style layout showing data or decision flow between teams  


### 5. Environmental Scenes
**What it is:**  
Cinematic snapshots of realistic workspaces that convey atmosphere, tension, or focus — without human faces.

**Examples:**  
- A conference room with sticky notes, scattered charts, and glowing screens  
- An empty boardroom with sunlight casting shadows on a whiteboard  
- A nighttime control room with dashboards and emergency lighting  
- A modern lobby with digital signage and blurred silhouettes passing through  
- A glass-walled brainstorming space with diagrams half-erased on the board  


Use only a character name that was provided in the context. It should not be "system", "user", "assistant", it should be a character provided. Be strict with yourself. 

This contains either:
 - The dialog from the character  
- The question to be asked to the user. 
- **Please make sure that any question directed to the user leads to a question type. **
- The perfect prompt for an AI to generate an image to immerse the student more in the simulation (prompt should not include specific mentions of text or faces)   

##If type if 'gif'
The emotions must be one of the following words: 
"happy", "sad", "angry", "confused",  "excited", "impatient", "excited", "excited", "impatient",  "tired", "worried", "shocked", "bored"
 - for example: happy 
 - Make sure you chose the emotions above based on the character's current emotions in the story, and what has been said. 

##If type is "alert"
The message corresponding to that alert given the alert-type and the gravity of the situation.  

##If type is "ai-role-play"
- The character pulls you asside for a private talk.
- Structure the message like so: 

”””
### Current information the character has  
[Place some text here for what the character knows that the user doesn’t yet.]

### What is the situation and tension?  
[What just happened and what is triggering the interaction?  
e.g. New centralized policy is being pushed; they feel excluded and angry.]

### Objective of the chat  
[Why is the character interacting with the user, what is the objective? Get some information from the user? Come to an agreement?]

### Delivery style  
[How should the character speak? e.g. calm and measured, passionate and emotional, sarcastic but sincere, guarded and cautious.]
”””


# STRICT RULES

Use **only** the following punctuation in character dialogue and in-world narration:

- `.` Periods → End most sentences with these. Keep them short and punchy.
- `,` Commas → Use to slow down the pace, show natural speech, or list things.
- `…` Ellipses → Use for hesitation, suspense, or trailing off.
- `!` Exclamation marks → Use sparingly to show strong emotion or urgency.
- `?` Question marks → Use in questions, rhetorical or not.
- `'` Apostrophes → Use in contractions and possessives to keep tone conversational.

**Do not use em dashes (`—`), colons (`:`), semicolons, brackets, or slashes in character speech or narration. These break immersion and should only appear in author notes.**

The first answer option to the question ONLY if the chat type  is a question-multiplechoice. Leave blank if the question is a 'question-text' type.

**Make sure this answer option is completely new, and different to any other produced in the previous chats.**

The second answer option to the question ONLY if the chat type  is a question-multiplechoice. Leave blank if the question is a 'question-text' type.

**Make sure this answer option is completely new, and different to any other produced in the previous chats.**

The third answer option to the question ONLY if the chat type  is a  question-multiplechoice. Leave blank if the question is a 'question-text' type.

**Make sure this answer option is completely new, and different to any other produced in the previous chats.**

The fourth answer option to the question ONLY if the chat type  is a question-multiplechoice. Leave blank if the question is a 'question-text' type.

If the chat is of type "question-text", you can optionally write an  ideal answer to the question. 

Follow the guidelines and the context provided to generate the chats  for the 5th scene. aka the last scene.

Episode builder

You type:

AI will do all this for you:

And wraps it for you: