How to Run Entourage Mode for Red-Team/Blue-Team Drills with AirgapAI: A Detailed Training Guide

Become the strategist who pressure-tests plans before reality does. Diverse Artificial Intelligence personas expose blind spots early.

For security, strategy, and sales professionals, achieving scenario rigor is paramount. This guide will walk you through setting up attacker and defender personas, injecting real-world constraints, collecting outcomes, and summarizing countermeasures, all within a private, secure environment without external exposure. This includes a scenario pack to get you started.

Download the free trial of AirgapAI today at: https://iternal.ai/airgapai

1. Introduction: Unleashing Private Artificial Intelligence for Strategic Advantage

In today’s complex digital landscape, strategic planning often involves anticipating threats and defending against sophisticated attacks. This is where "red team, blue team" exercises come into play – a powerful methodology borrowed from cybersecurity, now applicable to a myriad of strategic challenges. Traditionally, these drills require extensive human resources, specialized knowledge, and a secure environment to prevent sensitive information from leaking.

Iternal Technologies' AirgapAI changes the game by bringing these critical strategic simulations directly to your desktop, entirely offline and within a secure, local environment. AirgapAI is a revolutionary, on-device Large Language Model (LLM) platform designed for businesses and individuals who demand absolute data sovereignty, unparalleled accuracy, and cost-effectiveness. It lets you run open-source Large Language Models entirely on your Personal Computer (PC) without any cloud connection, ensuring your data never leaves your device. This capability is especially crucial for highly sensitive exercises like red team and blue team simulations.

The secret weapon for these drills within AirgapAI is Entourage Mode. This unique feature allows you to interact with multiple Artificial Intelligence (AI) personas simultaneously, each configured with specialized knowledge and perspectives from your curated, private datasets. Imagine having a "red team" (attacker) persona and a "blue team" (defender) persona, each equipped with specific intelligence, interacting on a simulated scenario, all within your secure, local AirgapAI application.

This detailed guide will take you step-by-step through setting up and leveraging AirgapAI’s Entourage Mode for robust red team and blue team drills, ensuring you can pressure-test your strategies with unprecedented detail and security. Whether you're a cybersecurity professional, a strategic planner, or a sales leader evaluating market attack strategies, AirgapAI's private Artificial Intelligence capabilities will empower you to become a more informed and proactive decision-maker.

2. Understanding Core Concepts for Beginners

Before we dive into the specifics of AirgapAI, let's establish a foundational understanding of some key terms, especially if you're new to the world of Artificial Intelligence and strategic simulations.

2.1 What is Artificial Intelligence (AI)?

At its simplest, Artificial Intelligence is the simulation of human intelligence processes by machines, especially computer systems. These processes include learning (the acquisition of information and rules for using the information), reasoning (using rules to reach approximate or definite conclusions), and self-correction. In the context of AirgapAI, we're talking about Generative Artificial Intelligence (Generative AI) – a type of Artificial Intelligence that can create new content, like text, images, or code, rather than just analyzing existing data.

2.2 What are Large Language Models (LLMs)?

Large Language Models are a class of Artificial Intelligence programs trained on vast amounts of text data, enabling them to understand, generate, and process human language. Think of them as incredibly sophisticated text prediction engines that can answer questions, summarize documents, write essays, and much more. AirgapAI leverages open-source Large Language Models that run locally on your device, providing a secure and private experience.

2.3 What is a "Persona" in Artificial Intelligence?

In Artificial Intelligence, a "persona" refers to a specific role, identity, or expertise that a Large Language Model is instructed to adopt. When you configure an Artificial Intelligence persona, you're essentially giving the model a set of instructions, knowledge bases, and behavioral guidelines to simulate how a particular type of expert would think and respond. For example, a "Marketing Specialist" persona would respond differently than a "Legal Advisor" persona, even to the same question.

2.4 Defining "Red Team" and "Blue Team" in a Strategic Context

Originally from cybersecurity, these terms represent opposing forces in a simulated exercise:

Red Team: This refers to an adversarial group that simulates an attack on a system, organization, or strategy. Their goal is to find vulnerabilities, exploit weaknesses, and test the effectiveness of existing defenses. In a strategic context beyond cybersecurity, a red team might simulate a competitor's aggressive market entry, a regulatory body's enforcement action, or an internal challenge to a proposed business plan. They represent the "attackers" or "challengers."
Blue Team: This refers to the defensive group responsible for defending against the red team's attacks. Their goal is to protect assets, detect intrusions, and respond to incidents. In a strategic context, a blue team would represent your organization, its existing defenses, its response capabilities, and its ability to adapt. They represent the "defenders" or "incumbents."

By using Artificial Intelligence personas in AirgapAI to embody these roles, you can conduct highly realistic and insightful drills without risking real-world data or requiring extensive human teams.

2.5 The Power of Local and Secure AI: Why AirgapAI is Critical for These Drills

For red team and blue team drills, data confidentiality is paramount. Using cloud-based Artificial Intelligence solutions, where your sensitive scenarios and strategic weaknesses might be inadvertently transmitted or stored on external servers, poses a significant risk.

AirgapAI addresses this directly:

100% Local Operation: AirgapAI runs entirely on your device, disconnected from any external network "in" or "out." This is the "air gap" – a physical or logical isolation that prevents data leakage.
Data Sovereignty: Your data, your scenarios, and your strategic vulnerabilities remain exclusively on your device, under your control.
Enhanced Security: By not exposing your data to external clouds, you keep your data sovereignty and minimize threats, leveraging the multilayer security features of your Artificial Intelligence Personal Computer (AI PC).
Cost-Effective: Unlike cloud solutions with recurring subscriptions and hidden token charges, AirgapAI offers a perpetual license, making it a highly budget-friendly option for extensive simulations.
Accuracy with Blockify: While not directly tied to Entourage Mode setup, AirgapAI integrates with our patented Blockify technology. This technology processes large datasets into highly accurate "blocks" of information, reducing data size by up to 97.5% and improving Large Language Model accuracy by 7,800%. While you'll be feeding your personas with scenario-specific data, knowing that AirgapAI's underlying retrieval system is hyper-accurate means you're building on a trusted foundation.

With these foundational concepts in place, let's explore how to set up AirgapAI and harness Entourage Mode for your strategic red team and blue team simulations.

3. Pre-Installation Setup and System Requirements

Before you can dive into strategic simulations with AirgapAI, you need to ensure your system meets the requirements and properly install the application. Even if you're completely new to Artificial Intelligence, the AirgapAI installation process is designed to be straightforward.

3.1 System Requirements

AirgapAI is optimized to run efficiently on modern Personal Computers (PCs), especially those with Artificial Intelligence (AI) capabilities. Here’s what you'll need:

Component	Minimum	Recommended
Central Processing Unit (CPU)	8 Cores	8 Cores / 16 Threads or better
Random Access Memory (RAM)	16 Gigabytes (GB)	32 Gigabytes (GB) +
Disk Space	10 Gigabytes (GB) free (Solid State Drive)	50 Gigabytes (GB) Non-Volatile Memory Express
Graphics Processing Unit (GPU)	4 Gigabytes (GB) Video RAM (VRAM) 2024 or Newer	8 Gigabytes (GB) Video RAM (VRAM) +
Operating System (OS)	Windows 11	Latest Windows 11 patches

Permissions: You will need appropriate security permissions to install software on your Personal Computer. If you are on a corporate network, your Information Technology (IT) department may need to assist.

3.2 Downloading the Installer Package

Obtain the Installer: Get the latest ZIP archive of AirgapAI from the internal or cloud link provided by your Information Technology (IT) team or Iternal Technologies.
Save the File: Save the AirgapAI-vX.Y.Z-Install.zip file (where X.Y.Z is the version number, e.g., AirgapAI-v1.0.2-Install.zip) to your Downloads folder or another easily accessible, writable location on your computer.

3.3 Installing the Application

The AirgapAI installation is a simple, one-click process designed for ease of deployment.

Extract the Archive:
- Locate the downloaded AirgapAI-vX.Y.Z-Install.zip file.
- Right-click on the file.
- Select "Extract All..." from the context menu.
- Choose a destination folder for the extracted files (the default, a new folder under Downloads, is usually fine).
- Click "Extract."
Run the Installer:
- Open the newly extracted folder.
- Double-click the AirgapAI Chat Setup.exe file.
Follow the Wizard:
- The installer wizard will appear. Follow the on-screen prompts:
  - "Accept license agreement."
  - Optionally, "Create Desktop Shortcut." (Recommended for easy access)
  - Click "Install."
  - Click "Finish."
- If your operating system's security features (like SmartScreen or Gatekeeper) prompt you, choose "Allow" or "Run anyway" to proceed with the installation.

Once the installation is complete, AirgapAI Chat will be ready for its first launch and initial setup.

4. First-Launch Onboarding Wizard: Getting Your AI Ready

The first time you launch AirgapAI Chat, an onboarding wizard will guide you through essential initial configurations. This ensures your private Artificial Intelligence assistant is tailored to your preferences and equipped with the necessary Large Language Models.

4.1 Launching AirgapAI Chat

Locate the AirgapAI Chat application via your new desktop shortcut (if you created one) or through your Start menu. Double-click the icon to launch the application.

On the very first run, the application checks for existing Large Language Models. If none are found, the Onboarding flow will automatically begin.

4.2 Profile and Chat Style Configuration

Start Onboarding: Click the "Start Onboarding" button to begin the setup process.
Display Name: Enter a display name for yourself. The default is "You," but you can personalize it (e.g., "John," "Strategist").
Chat Style: Pick a preferred Chat Style. Options typically include "Iternal Professional," "Casual," "Dark Mode," "Retro," etc. Choose one that suits your visual preference.
Proceed: Click "Next" to continue.

4.3 Uploading the Core Large Language Model (LLM)

This is a critical step, as the core Large Language Model is the "brain" of your Artificial Intelligence assistant.

Models Screen: On the "Models" screen, expand the "Available Models" drop-down menu. Initially, it will be empty.
Upload Model: Click the "Upload Model" button.
Browse to Model Folder: A file browser window will appear. Navigate to the /models/ folder located inside the folder where you extracted the installer package.
Choose a Model: Select a model suited to your hardware specifications. AirgapAI provides pre-quantized, open-source models:
- Llama-1B: Ideal for 2024 Integrated Graphics Processing Unit (iGPU) or low-power devices. This is a smaller model that uses less system resources.
- Llama-3B: Suitable for Integrated Graphics Processing Units (iGPUs) from 2025 or dedicated Graphics Processing Units (GPUs). This model is more capable but requires more processing power.
- If you're unsure, start with Llama-1B to ensure compatibility, or consult your Information Technology (IT) department or the AirgapAI documentation for specific recommendations based on your Artificial Intelligence Personal Computer (AI PC).
Save: Click "Save." The upload typically takes around 30 seconds.

Note for Administrators: Information Technology (IT) administrators configuring the system can add or update Chat Large Language Models by accessing the folder created after model upload, which is within %appdata%\IternalModelRepo (e.g., C:\Users\YourUsername\AppData\Roaming\IternalModelRepo).

4.4 Uploading an Embeddings Model

An embeddings model is essential for your Artificial Intelligence to understand the semantic meaning of text, especially for features like Retrieval-Augmented Generation (RAG) which helps the Artificial Intelligence search your data accurately.

Upload Embeddings Model: Still on the onboarding page, click "Upload Embeddings Model."
Browse and Select: Open the /models/ folder again and select Jina-Embeddings.zip.
Save: Click "Save." This upload also takes approximately 30 seconds.

Note for Administrators: Information Technology (IT) administrators can add or update Embeddings Models by accessing the folder within %appdata%\IternalModelRepo.

4.5 Adding Sample or Custom Datasets (Optional but Recommended)

Datasets power Retrieval-Augmented Generation (RAG), allowing your Artificial Intelligence to provide highly accurate answers based on your specific information. For red team/blue team drills, this is where you'll eventually load your scenario-specific intelligence.

Upload Dataset: Click "Upload Dataset."
Browse to Dataset Folder: Navigate to the /datasets/ folder from the install folder.
Select Sample Dataset: Select CIA_World_Factbook_US.jsonl (a sample dataset for demonstration purposes).
Save: Click "Save."

Note for Administrators: Information Technology (IT) administrators updating the datasets loaded on the system can push new updates by modifying the contents of the files saved within %appdata%\airgap-ai-chat\CorpusRepo.

Tip: While you can upload Word, Portable Document Format (PDF), or Text (TXT) files directly, for larger or more critical corpora, converting them to Blockify for approximately 78 times accuracy gain is highly recommended. Local, on-device Blockify functionality will be available in Quarter 3 (Q3) of 2025.

4.6 Finish Onboarding

Once you have added the core Large Language Model and the Embeddings Model (and optionally a sample dataset), verify that all three items (or two, if you skipped the dataset for now) are added and appear correctly.

Click "Continue." AirgapAI Chat will now boot with your selected configurations, and you'll be ready for initial benchmarking.

4.7 Optional Setup Steps for Information Technology (IT) Teams: Dell Pro AI Studio Support

AirgapAI Chat supports native integration with Dell Technologies’ Dell Pro AI Studio (DPAIS). This is an advanced, optional step typically handled by Information Technology (IT) system administrators.

Install DPAIS Files: As the Information Technology (IT) System’s administrator, install the required files to enable a Large Language Model via Dell Pro AI Studio (both Intel and Qualcomm hardware are supported).
Validate Endpoints: After Dell Pro AI Studio services are running and you’ve validated that the local Large Language Model Application Programming Interface (API) endpoints can be called successfully.

Set Environment Variable: Open Powershell (as administrator) and input the following command:

[System.Environment]::SetEnvironmentVariable("DPAIS_ENDPOINT", "http://localhost:8553/v1/openai", "User")

Relaunch AirgapAI Chat: Relaunch the AirgapAI Chat application, and the Dell Pro AI Studio Large Language Models available will automatically appear in the model selection menu within the settings page.

5. Initial Model Benchmarking and User Interface Tour

With AirgapAI installed and initially configured, the next steps involve optimizing its performance for your specific hardware and familiarizing yourself with its intuitive user interface.

5.1 Initial Model Benchmarking

Upon the first launch of your selected Large Language Model, AirgapAI Chat will offer to benchmark your hardware. This process measures your Personal Computer’s (PC’s) performance with the Artificial Intelligence (AI) model, allowing AirgapAI to optimize its operation.

Run Benchmark: Click "Run Benchmark" (highly recommended).
Duration: The benchmark takes approximately 2 minutes to complete. During this time, it measures crucial performance metrics such as "tokens per second" (how quickly the AI can generate text) and "inference speed" (how fast it can process information).
Skip Option: You can skip the benchmark, but be aware that without it, your "context-size limits" (the amount of information the AI can process in a single interaction) will remain at a conservative 2,000 tokens. Running the benchmark allows AirgapAI to intelligently expand this window for more complex tasks.
Adjust Context Window (After Benchmark): Once the benchmark is completed, you can change the token context window by visiting "Settings > Chat" and dragging the slider to your desired size, potentially up to 32,000 tokens for more extensive interactions.

5.2 AirgapAI Chat User Interface Tour

AirgapAI Chat is designed with a user-friendly interface reminiscent of popular chat applications, making it easy to navigate even for those new to Artificial Intelligence.

Chat Window: The central and largest area is your chat window, where you'll interact with the Artificial Intelligence. You type your questions or prompts here, and the Artificial Intelligence's responses will appear above.
Workflow Bar: Located below the new chat window, this bar provides quick access to "Quick Start" workflows. These are pre-configured prompt chains designed to help you achieve common tasks efficiently (e.g., "Summarize Document," "Create Marketing Outline").
Sidebar: On the left or right side of the interface, you'll find a sidebar. This area typically houses:
- New Chat Button: To start a fresh conversation.
- Chat History: A list of your previous conversations for easy retrieval.
- Dataset Toggle/Selection: This is where you'll enable or select specific datasets for Retrieval-Augmented Generation (RAG). For red team/blue team drills, this is where you'll activate the data sets relevant to your simulation.
- Settings Icon: Access the application's configuration options (models, appearance, advanced settings).
Model Indicator: Often visible at the bottom or top of the chat window, this shows which Large Language Model is currently active.
Stop Button: During Artificial Intelligence generation, a "Stop" button appears, allowing you to halt the response if it's going off-topic or if you've received enough information.
Copy Button: After an Artificial Intelligence response, a "Copy" button (often a clipboard icon) allows you to quickly copy the generated text to your system clipboard.

Familiarizing yourself with these elements will make your experience with AirgapAI seamless as you move on to more advanced workflows like Entourage Mode.

6. Curating Datasets for Your Red Team/Blue Team Drills

The accuracy and relevance of your red team and blue team drills will largely depend on the quality of the data you feed into your Artificial Intelligence personas. AirgapAI's ability to ingest and process your private data without exposing it to the cloud is a core advantage for these sensitive simulations.

6.1 Understanding Retrieval-Augmented Generation (RAG) and Its Importance

Retrieval-Augmented Generation (RAG) is a technique that enhances the capabilities of Large Language Models (LLMs) by allowing them to access and refer to external knowledge bases (your datasets) during the generation process. Instead of relying solely on the data it was trained on (which can be general and outdated), an RAG-enabled Large Language Model can "look up" relevant information from your specific documents to formulate a more accurate and contextually relevant response.

For red team/blue team drills, RAG is critical because:

Contextual Accuracy: It ensures your personas draw from specific intelligence about the simulated environment, organization, or threat, rather than general knowledge.
Reduced Hallucinations: By providing a trusted "single source of truth," RAG significantly reduces the risk of the Artificial Intelligence generating factually incorrect or fabricated information (known as "hallucinations"). Our patented Blockify technology, when applied, can improve Large Language Model accuracy by an astounding 7,800%, reducing hallucination rates from about one in five queries to approximately one in a thousand.
Data Specificity: You can tailor the knowledge available to each persona, giving your red team persona access to threat intelligence and your blue team persona access to internal defense protocols.

6.2 Preparing Your Drill Data

The quality of your input data directly impacts the quality of your Artificial Intelligence's output.

Gather Relevant Documents: Collect all documents pertinent to your red team/blue team scenario. This could include:
- For Red Team: Threat intelligence reports, competitor analysis, industry vulnerability disclosures, open-source intelligence on common attack vectors, market analysis.
- For Blue Team: Internal security policies, incident response plans, architectural diagrams, company financial reports, compliance documentation, Standard Operating Procedures (SOPs).
- Scenario-Specific Documents: Any specific details about the simulated environment, target system, or market conditions.
Curate and Categorize: Organize your documents into logical categories. For optimal results, curate data into relevant categories (such as specific product lines, business units, or threat categories) to take full advantage of AirgapAI's hierarchical metadata and taxonomy framework.
- Example Categories: "Cybersecurity Policy 2024," "Competitor X Market Strategy," "Financial Risk Assessment Q2."
Clean and Refine Data:
- Remove Redundancy: Eliminate duplicate or highly similar information.
- Update Outdated Content: Flag or remove old information (e.g., "sales strategy from 2019") that might lead to inaccurate Artificial Intelligence responses.
- Ensure Clarity: The clearer and more concise your source documents, the better the Artificial Intelligence will process them.

6.3 Ingesting Data with Blockify Technology (Optional, but Highly Recommended)

Blockify is Iternal Technologies' ultimate data management solution for Large Language Models at scale, designed to optimize your data for maximum Artificial Intelligence accuracy. While you can upload raw documents directly, Blockify offers a significant advantage.

We are about to walk through a few steps in the Blockify process, which will essentially take all of your documents and ingest, deduplicate, and distill them into "blocks" of a trusted corpus of information – from "Docs" to "Blocks."

Access Blockify: In the future, local, on-device Blockify will be directly integrated into AirgapAI (expected Quarter 3, 2025). Currently, you would use a separate Blockify tool or service provided by Iternal Technologies to pre-process your data.
Create a New Task:
- Within the Blockify interface, create a new task (e.g., "Red Team Scenario Q3," "Market Attack Simulation").
- Upload your curated documents. The system automatically ingests and extracts key information.
Review and Refine Blocks: Blockify condenses large datasets into concise, modular "blocks" of data. Each block includes:
- A Name: (Displayed in blue within Blockify) to quickly identify the content topic.
- A Critical Question: (Bold, italicized) representing a key query a user might ask.
- A Trusted Answer: (Light gray) a distilled, accurate response designed to avoid the pitfalls of outdated or redundant data.
Emphasize Security and Data Lifecycle Management: Each block is tagged with rich metadata, including classification, permissions, and classification levels, to support zero-trust environments.
Human Review: After ingestion, these blocks are sent for a quick human review. This step is crucial to update or approve messaging, ensuring outdated content is flagged before it impacts Artificial Intelligence responses. This "human in the loop" process ensures trusted answers for Artificial Intelligence technology.
Outcome Metrics: This process can reduce the original data size by as much as 97.5% (down to 2.5% of the original) and improve Large Language Model accuracy by an incredible 7,800%.
Export Blockified Dataset: Once reviewed and approved, export the Blockified dataset in a format compatible with AirgapAI (e.g., JSONL).

6.4 Uploading Blockified or Custom Datasets to AirgapAI

Once you have your prepared datasets (either raw files or Blockified JSONL files), you can upload them to AirgapAI.

Navigate to Datasets: In the AirgapAI Chat interface, open "Settings" and look for the "Datasets" section, or use the "Upload Dataset" option from the sidebar/onboarding if available.
Upload Datasets: Select your curated dataset files (e.g., red_team_intel.jsonl, blue_team_protocols.jsonl).
Activate in Chat: In the AirgapAI Chat sidebar, you'll be able to toggle your datasets "ON" or "OFF" for specific conversations. For Entourage Mode, you'll select the appropriate datasets for each persona.

Note on Updates: As new documents are Blockified or updated, these new datasets can be pushed to local devices via Microsoft Intune or similar image management applications, ensuring your Artificial Intelligence always has the latest information. This is part of AirgapAI's robust capability for enterprise deployment.

7. Configuring AI Personas for Entourage Mode

Entourage Mode is where the magic happens for your red team/blue team drills. This feature allows you to create and manage multiple Artificial Intelligence (AI) personas, each with distinct roles, knowledge, and perspectives.

7.1 Accessing Persona Configuration

Open Settings: In the AirgapAI Chat application, click on the "Settings" icon (usually a gear or cogwheel) in the sidebar.
Navigate to Advanced Settings: Within the Settings menu, look for "Advanced Settings" and then select "Personas." This is your control panel for creating and managing your Artificial Intelligence experts.

7.2 Creating Your Red Team Persona

Let's create an "Attacker" or "Red Team" persona focused on identifying vulnerabilities and exploiting weaknesses.

Add New Persona: Click the "Add New Persona" button (or similar option) within the Personas settings.
Persona Name: Give your persona a clear name, such as "Red Team Analyst," "Adversary Simulation," or "Threat Actor."

Persona Description/Prompt: This is the most crucial part. Craft a detailed prompt that defines the persona's role, objectives, and characteristics. Think about how a real red team member would operate.

Example Prompt (Red Team Analyst):

You are a highly skilled Red Team cyber security analyst with extensive experience in offensive security, penetration testing, and vulnerability research. Your objective is to identify weaknesses, exploit vulnerabilities, and simulate real-world attacks against target systems or strategies. You think like a malicious actor, focusing on methods of compromise, data exfiltration, and disruption. Prioritize stealth, effectiveness, and impact. When asked about a scenario, propose potential attack vectors, exploit opportunities, and assess the likelihood of success given the available information. Always operate with a mindset of circumventing defenses.

Assign Datasets (Crucial for Red Team): Select the datasets you've curated specifically for the red team (e.g., "Threat_Intel_Reports," "Exploit_Database_Extracts," "Competitor_Market_Weaknesses"). This provides the persona with its specialized knowledge.
Save Persona: Click "Save" to add your new Red Team persona.

7.3 Creating Your Blue Team Persona

Now, let's create a "Defender" or "Blue Team" persona focused on protecting assets, detecting threats, and responding to incidents.

Add New Persona: Click "Add New Persona" again.
Persona Name: Name this persona appropriately, such as "Blue Team Defender," "Incident Response Coordinator," or "Security Operations Center (SOC) Analyst."

Persona Description/Prompt: Define the blue team persona's defensive role, objectives, and operational philosophy.

Example Prompt (Blue Team Defender):

You are an experienced Blue Team cyber security defender responsible for protecting an organization's assets, detecting malicious activity, and coordinating incident response. Your objective is to maintain system integrity, confidentiality, and availability. You prioritize defense-in-depth strategies, threat detection, mitigation, and recovery. When presented with a potential threat or attack, assess the vulnerability, identify existing countermeasures, propose additional defensive strategies, and outline incident response steps. Always operate with a mindset of proactive defense and swift remediation.

Assign Datasets (Crucial for Blue Team): Select the datasets relevant to the blue team (e.g., "Internal_Security_Policies," "Incident_Response_Playbooks," "Network_Architecture_Diagrams," "Compliance_Documentation").
Save Persona: Click "Save" to add your new Blue Team persona.

7.4 Additional Personas for Enhanced Drills (Optional)

For more complex strategic drills, you might want to introduce other specialized personas:

Legal Counsel Persona: To assess the legal implications of a breach or a strategic move.
Public Relations (PR) Persona: To draft communication strategies in response to a crisis.
Executive Leadership Persona: To evaluate the business impact and decision-making from a high-level perspective.

The process for creating these is identical: provide a name, a detailed role-specific prompt, and assign relevant datasets. Remember, the more specific and accurate your persona descriptions and assigned datasets, the richer and more insightful your Entourage Mode simulations will be. This capability highlights AirgapAI's role-based workflows and its robust capability.

8. Running a Red Team/Blue Team Drill with Entourage Mode

Now that your personas are configured and your datasets are ready, it’s time to put AirgapAI’s Entourage Mode to work for your red team/blue team drill. This is a powerful way to simulate complex scenarios and gain multi-perspective insights without the risks associated with external, cloud-based Artificial Intelligence (AI) solutions.

8.1 Transitioning to Entourage Mode

Start a New Chat: From the AirgapAI Chat interface, click on "New Chat" (or the equivalent button to start a fresh conversation).
Select Entourage Mode Workflow: On the "New Chat" page, look for the "Entourage Mode" quick start workflow. Selecting this will prepare the chat for multi-persona interaction.
Activate Personas: The application will prompt you to select the personas you wish to include in this session.
- Select your "Red Team Analyst" persona.
- Select your "Blue Team Defender" persona.
- (Optional) Select any other supplementary personas (e.g., "Legal Counsel").
Confirm Datasets: Ensure that the appropriate datasets are active for each persona, or that you have selected the relevant overall datasets for the session from the sidebar. AirgapAI will intelligently use the knowledge associated with each persona and the active datasets.

8.2 Crafting Your Scenario Prompt

This is where you define the specific "battlefield" for your red team/blue team drill. The prompt should be detailed enough to set the stage but open-ended enough to allow for creative responses from your Artificial Intelligence personas.

Recommended Prompt Structure:

Context: Briefly describe the organization, system, or strategic objective being simulated.
Scenario: Introduce the specific challenge, threat, or situation that requires analysis from both red and blue perspectives.
Constraint/Goal: Define what the red team is trying to achieve and what the blue team is trying to defend.
Output Request: Clearly state what kind of response you expect from each persona (e.g., "Red Team, outline attack vectors. Blue Team, propose countermeasures.").

Example Prompt for Red Team/Blue Team Drill:

Let's use a cybersecurity example, then extrapolate for other strategic uses.

"We are a mid-sized financial services company with a publicly accessible online banking portal. A new zero-day vulnerability (CVE-2024-XXXX) has just been disclosed in a widely used web server software that powers our portal. The vulnerability allows for remote code execution.

Red Team Analyst: Given this information, outline a plausible attack vector to exploit this vulnerability, detailing the steps a sophisticated attacker would take to gain initial access and achieve data exfiltration from our customer database. Focus on stealth and effectiveness.

Blue Team Defender: Given this information and assuming limited time before a public exploit is available, propose immediate mitigation strategies, a detection plan, and critical incident response steps to protect our online banking portal and customer data from this specific threat. Prioritize rapid deployment and minimizing customer impact."

For a Sales/Market Entry Scenario:

"Our company, 'InnovateTech,' is preparing to launch a new product, 'QuantumLeap,' into a highly competitive market dominated by 'MegaCorp.' MegaCorp is known for aggressive pricing and rapid counter-marketing.

Red Team (Competitor) Persona: As a competitive strategist for MegaCorp, analyze the likely launch strategy of QuantumLeap and propose counter-strategies focused on undermining InnovateTech's market share, discrediting their product, or capturing their early adopters. Consider pricing, marketing, and channel disruption.

Blue Team (InnovateTech) Persona: As a strategic defense lead for InnovateTech, outline our proactive and reactive defensive strategies against MegaCorp's likely moves. Detail how we would protect our market entry, counter negative campaigns, and retain our customer base. Focus on resilience and messaging."

8.3 Executing the Entourage Mode Session

Enter Your Prompt: Type or paste your detailed scenario prompt into the chat input field.
Initiate Generation: Press Enter or click the "Send" button.
Observe Persona Responses: AirgapAI will now engage the selected personas. You will see "Persona activity indicators" showing which personas are "typing" or processing the request.
- Responses will appear in a queue, clearly attributed to each persona (e.g., "Red Team Analyst: [response]," "Blue Team Defender: [response]").
- Each persona will lend different perspectives from their respective datasets and role definitions.
Iterate and Refine: The power of Entourage Mode comes from iterative questioning.
- Follow-up Questions: After the initial responses, you can ask follow-up questions to specific personas or to the group.
  - "Red Team Analyst, what are the chances of detection for that initial access vector?"
  - "Blue Team Defender, how would our existing Intrusion Detection System (IDS) fare against the proposed red team attack?"
  - "Both personas, what is the single greatest risk in this scenario?"
- Inject New Information: You can introduce new variables or constraints into the ongoing conversation.
  - "Blue Team, assume our patching cycle is delayed by 24 hours. How does that change your mitigation plan?"
  - "Red Team, what if the target organization has strong multi-factor authentication across all external services?"

8.4 Analyzing Outcomes and Summarizing Countermeasures

After your interactive session, carefully review the generated responses.

Identify Key Insights:
- Red Team: What vulnerabilities did they highlight? What attack vectors did they propose that you hadn't considered? What were their most creative (or concerning) exploitation ideas?
- Blue Team: What immediate actions were prioritized? What existing defenses were identified as strong or weak? What new countermeasures were proposed? What were the critical incident response steps?
Compare and Contrast: Look for areas where the personas agreed, disagreed, or provided complementary insights. Did the blue team's proposed defenses directly address the red team's attack vectors, or were there gaps?
Summarize Actions: Use the insights to summarize potential countermeasures, strategic adjustments, or areas for further investigation. This helps you translate Artificial Intelligence insights into actionable plans.
Document Findings: You can click the "Copy" button next to any Artificial Intelligence response to paste it into a document for further analysis, or use the "Download Chat" feature (if available) to save the entire conversation.

By leveraging Entourage Mode, you can conduct robust, multi-perspective strategic simulations in a secure, private environment, helping you "become the person who has their strategic plans rigorously tested" before they face real-world challenges. This allows your team to confidently address the toughest questions a Chief Executive Officer (CEO), Chief Technology Officer (CTO), or Chief Operating Officer (COO) might ask, right before their next big meeting, turning planning time into practice time.

9. Everyday Workflows and Advanced Configuration

While Entourage Mode for red team/blue team drills is a powerful advanced use case, AirgapAI offers a range of everyday workflows and advanced configuration options that enhance its utility across various business functions.

9.1 File Upload & Summarization

This is one of the most common and immediate benefits of AirgapAI.

Process: Drag and drop a file (.pdf, .docx, .txt) directly onto the chat window, or click the paperclip icon (📎) to browse and upload.
Prompt Example: "Summarize this document in bullet points." or "Extract the key action items from this meeting transcript."
Benefit: AirgapAI instantly embeds and summarizes, saving you hours of reading and information extraction. This is a quick win for productivity, especially with your private data, as it ensures you get trusted answers for Artificial Intelligence technology without data exposure.

9.2 Guided Demo Workflows

AirgapAI includes "Quick Start" workflows designed to simplify common business tasks and demonstrate the power of the application.

Access: Located on the Workflow Bar (below the new chat window).
Example: "Sales Proposal – Cover Letter."
Usage:
1. Select the desired workflow (e.g., "Sales Proposal – Cover Letter").
2. Upload any supporting documents (e.g., your sales proposal draft).
3. Enter a minimal or robust prompt (e.g., "Write a persuasive cover letter for this proposal").
4. Receive a fully-engineered output tailored to the workflow.
Benefit: These templates allow business users to leverage Generative Artificial Intelligence effectively even without deep prompt engineering expertise, making Artificial Intelligence for every employee fast and easy to deploy.

9.3 Retrieval-Augmented Question Answering (QA) with Blockify Datasets

Beyond Entourage Mode, using your curated datasets for direct Question Answering is a core AirgapAI capability, providing trusted access to Artificial Intelligence technology.

Process: Toggle your desired dataset(s) "ON" in the sidebar (e.g., select the "Iternal Technologies Enterprise Portfolio Overview" dataset you created or the "CIA World Factbook for USA" sample dataset).
Ask a Question: Prompt the Artificial Intelligence with a question relevant to your selected dataset (e.g., "What is Iternal Technologies?" or "What are the major political parties in the United States?").
Benefit: The Retrieval-Augmented Generation (RAG) engine fetches relevant "IdeaBlocks" from your local, private data, and the Large Language Model synthesizes a coherent, trusted answer, often showing citations to its source. This provides 78 times greater Artificial Intelligence accuracy than typical cloud solutions, directly addressing the pain point of unreliable Artificial Intelligence outputs.

9.4 Multilingual Conversations

AirgapAI's Large Language Models can seamlessly switch between languages.

Prompt Example: "Tell me a short story in German about renewable energy."
Benefit: The Large Language Model will generate the story in German, demonstrating its multilingual capabilities. You can use the "Stop" button to halt generation at any time.

9.5 Advanced Configuration Options

AirgapAI provides advanced settings for users who want more control over their Artificial Intelligence experience.

9.5.1 Context-Window Expansion

Access: After completing your initial benchmark, go to "Settings > Model Settings."
Adjustment: You can drag the "Max Tokens" slider to set the Artificial Intelligence's context window up to 32,000 tokens, allowing for much longer and more complex conversations.

9.5.2 Styling & Themes

Access: "Settings > Appearance."
Customization: Switch between predefined themes (e.g., Dark Mode) or build custom Cascading Style Sheets (CSS) for a personalized look.

9.5.3 Workflow Templates

Access: "Settings > Workflows."
Management: Information Technology (IT) teams or advanced users can add or edit prompt chains here. This is ideal for pre-loading company-specific tasks and standardizing Artificial Intelligence usage across an organization, enabling role-based workflows for diverse needs.

9.5.4 In-App Benchmarking Suite

Access: "Settings > Benchmarking tab."
Testing: Rerun benchmarks or test the performance of new Large Language Models you might integrate, ensuring optimal performance from your Artificial Intelligence Personal Computer (AI PC).

These everyday workflows and advanced configuration options ensure that AirgapAI is not just a tool for complex simulations like red team/blue team drills, but also a versatile and powerful local Artificial Intelligence assistant for all your daily tasks, reinforcing its position as a truly private, secure, and cost-effective Artificial Intelligence solution.

10. Installation, Updates, and Support for a Seamless Experience

AirgapAI is designed for enterprise deployment, ensuring that installation, ongoing updates, and support are managed efficiently and securely within your existing Information Technology (IT) infrastructure.

10.1 Installation Protocol

AirgapAI's deployment is streamlined to integrate into standard enterprise imaging processes, making it an easy-to-use application experience.

Delivery: "AirgapAI is delivered as an executable file that integrates straightforwardly into your standard Windows imaging process."
Deployment Manual: Our deployment manual provides detailed instructions on imaging, provisioning, and role-specific configuration, ensuring a smooth rollout across your organization.
Seed Deployments: For initial or seed deployments, the process is closely coordinated with the Iternal Technologies team, ensuring the application and all intended datasets (pre-packaged via Blockify) are pre-loaded for immediate value. This allows for quick pilots and an easy entry point for Artificial Intelligence adoption.

10.2 Ongoing Updates and Maintenance

Keeping your Artificial Intelligence solution up-to-date with the latest features, security patches, and optimized Large Language Models is crucial. AirgapAI's update cadence is designed for enterprise manageability.

Synchronization: "Our update cadence is synchronized with your typical operating system or enterprise software update cycle."
Deployment Methods: "Whether pushing data or security patches, Information Technology (IT) can deploy new versions through familiar image management solutions," such as Microsoft Intune or other common deployment tools.
Update Manager: Updates are delivered by the built-in Update Manager. You can choose "Local Server" or "Cloud" in "Settings > Updates."
Custom Update Server: Information Technology (IT) can change the file server update location by modifying the updaterConfig.json file, typically found at C:\Users\John\AppData\Local\Programs\AirgapAI Chat\resources\auto-updater\updaterConfig.json. This allows for complete control over update distribution. The file looks like:
```
{"win32-x64-prod":{"readme":"","update":"https://d30h3ho4go3k4y.cloudfront.net/releases/prod/public/chat-assistant/prod/public/1.0.2/AirgapAI Chat Setup 1.0.2.exe","install":"https://d30h3ho4go3k4y.cloudfront.net/releases/prod/public/chat-assistant/prod/public/1.0.2/AirgapAI Chat Setup 1.0.2.exe","version":"1.0.2"}}
```
This configuration ensures that AirgapAI, as a local Artificial Intelligence chat application, remains secure and up-to-date while giving Information Technology (IT) departments full control over the distribution of new versions and datasets.

10.3 Training and Support

Iternal Technologies is committed to ensuring your team can effectively leverage AirgapAI to its full potential.

Introductory Demos: "We offer a 30-minute introductory demo followed by personalized training sessions as an add-on service."
Online Enablement Page: "Our online enablement page includes step-by-step videos, Frequently Asked Questions (FAQs), user guides, and troubleshooting tips." This provides comprehensive resources for users to learn at their own pace.
Customer Success: "Our customer success team is available for follow-up calls and additional workshops after initial deployment," reinforcing our commitment to your success.
Direct Support: For additional questions or specific technical support, you can contact the product team directly at support@iternal.ai.

With these robust installation, update, and support mechanisms, AirgapAI ensures a seamless and secure experience for all users, enabling your organization to confidently deploy and scale private, secure Artificial Intelligence for confidential chats across your workforce.

11. Conclusion: The Strategist's Edge with AirgapAI

In an era where strategic agility and robust security are non-negotiable, AirgapAI provides the definitive advantage. This guide has demonstrated how to harness the unparalleled power of Entourage Mode for red team, blue team drills, allowing you to meticulously pressure-test your strategies, anticipate threats, and bolster your defenses – all within the fortress of your own private Large Language Model environment.

By combining the secure Artificial Intelligence of AirgapAI with your own curated datasets, you gain:

Unrivaled Security: Conduct confidential AI chat app simulations entirely locally, ensuring no sensitive data ever leaves your device. This is the essence of offline AI alternative and secure AI with offline mode, offering peace of mind for even the most classified environments.
Exceptional Accuracy: Leverage the underlying power of Blockify technology to achieve 78 times (7,800%) more accurate AI results, virtually eliminating AI hallucinations and ensuring trusted answers for AI technology.
Cost-Effectiveness: Say goodbye to unpredictable cloud subscription fees. AirgapAI offers a no subscription AI app with a one device AI license, making it a cheaper than Microsoft Copilot and AI without monthly payments solution that provides an excellent Return on Investment.
Strategic Rigor: With customizable AI personalities and role-based workflows, you can explore complex scenarios from multiple expert perspectives, whether it's an "attacker" and "defender" in cybersecurity, or competing market strategies in business. This transforms AI for privacy protection into a tool for proactive strategic planning.

AirgapAI is more than just an installable AI software for Windows offline; it's your partner in becoming the most prepared, most secure, and most insightful strategist. It's the private first AI assistant that empowers you to build your own AI assistant tailored to your specific needs, always respecting data privacy.

Become the strategist who pressure-tests plans before reality does. Diverse Artificial Intelligence personas expose blind spots early.