How Does Sasha Work?

Understanding Sasha's Layered Intelligence System

The Three Context Layers

Sasha builds intelligence through three distinct layers of context. Each layer adds depth and specificity to create the "she just knows" experience that makes Sasha feel like an expert colleague rather than a basic AI assistant.

Global Context Layer

What it provides: The foundational intelligence that makes Sasha capable from day one.

This layer includes Sasha's built-in expertise:

• Professional Methodologies: Proven frameworks for analysis, documentation, research, and problem-solving developed over years of enterprise consulting
• Quality Standards: Enterprise-grade requirements for accuracy, verification, and presentation
• Domain Knowledge: Specialist expertise across legal, technical, financial, and operational domains
• Design Systems: Visual hierarchy principles, documentation standards, and communication patterns

The result: Even before learning anything about your organization, Sasha applies professional-grade methodologies to every task. When you ask for documentation, you get enterprise-quality output. When you ask for analysis, you get structured, verified insights.

Project Context Layer

What it provides: Your organization's specific knowledge, derived from your actual documents.

Sasha studies your enterprise documents - files on network drives, shared folders, uploaded documents - and derives understanding from them:

What Sasha Learns From Your Documents

How Knowledge Builds Over Time

Month 1: Sasha understands your document structure and basic terminology

Month 3: Sasha recognizes patterns across projects and clients

Month 6: Sasha predicts what approaches work for different situations

Year 1+: Sasha operates with deep institutional knowledge, making connections humans might miss

Output Style Context

What it provides: Control over how Sasha communicates with you.

Output styles shape Sasha's responses to match your needs:

• Explanatory Mode: Detailed explanations with reasoning and context - ideal for learning or complex decisions
• Concise Mode: Brief, actionable responses - ideal for experienced users who need quick answers
• Technical Mode: Precise, jargon-appropriate responses - ideal for specialist work
• Custom Styles: Your organization can define specific communication preferences

Why this matters: The same underlying intelligence can be presented differently depending on who's asking and what they need. A senior partner might want a brief summary. A new analyst might need the full explanation.

How Sasha Derives Knowledge

Unlike traditional AI that simply searches documents, Sasha understands them:

From Documents to Intelligence

Your Documents                    What Sasha Derives
──────────────                    ─────────────────
"Project X proposal              → "Similar projects typically
 for Client ABC,                    take 6-8 weeks, this client
 £45,000, 8 weeks"                  prefers phased approaches,
                                    pricing is competitive
"Similar proposal for               for the market"
 Client DEF, £52,000,
 10 weeks"

Pattern Recognition at Scale

Sasha doesn't just find documents - it synthesizes insights across your entire knowledge base:

• Pricing patterns: "What do we typically charge for this type of work?"
• Resource allocation: "Who has experience with similar projects?"
• Risk factors: "What went wrong in comparable situations?"
• Success patterns: "What approaches worked best for this client type?"

The "She Just Knows" Experience

When users say Sasha "just knows" things, they're experiencing the combination of all three context layers working together:

Example Query: "How should we price this retail transformation project?"

What Sasha considers:

1. Global Context: Standard transformation pricing methodologies, market benchmarks, value-based pricing frameworks

2. Project Context: Your organization's historical pricing for retail clients, similar project outcomes, this client's previous engagements with you

3. Output Style: Whether to provide a quick number or a detailed pricing analysis with reasoning

The Result: "Based on our 12 similar retail transformation projects, I'd recommend £65-75K. The Tesco project last year was £58K but ran over scope. The Sainsbury's engagement at £72K stayed on track. Given this client's complexity level, £70K positions us competitively while maintaining margin."

Why This Architecture Works

For Your Organization

• Institutional memory preserved: Knowledge doesn't leave when people do
• Expertise democratized: Junior staff access senior-level insights
• Consistency maintained: Same quality standards across all work
• Learning accelerated: New hires productive from day one

For Users

• Natural interaction: Ask questions like you would ask a knowledgeable colleague
• Contextual answers: Responses consider your specific situation
• Progressive depth: Get quick answers or detailed analysis as needed
• Continuous improvement: Sasha gets more helpful as it learns more

Getting Started

Sasha begins working immediately with global context. As you add your organization's documents and configure project contexts, the intelligence deepens.

Day 1: Enterprise-grade AI assistance with professional methodologies

Week 1: Understanding of your specific domain and terminology

Month 1: Recognition of patterns across your documents

Ongoing: Continuously improving institutional knowledge

Sasha - Transforming institutional knowledge into conversational intelligence