Project AI

Ask questions about your model in plain English and get instant answers about assumptions, calculations, and financial performance.

Project AI in Aire

Project AI is a conversational assistant built into Aire that helps you work with financial models faster and more intuitively. Instead of navigating through spreadsheets or tracing dependencies, you ask questions in plain English and get instant, context-aware answers.

The agent understands your entire model—every term, block, case, formula, and financial relationship. It can answer questions, explain calculations, identify risks, and help you make sense of complex project economics in seconds.

What It Does

Project AI can:

• Answer questions about terms, calculations, or assumptions
• Explain relationships between inputs and outputs across blocks
• Trace dependencies to understand how changes propagate through your model
• Compare scenarios by analyzing differences across case selections
• Identify risks by highlighting sensitive variables and potential issues
• Find specific data instantly without manual searching

The agent operates on your live model—answers always reflect current values and structure.

How It Works

Always Aware of Your Project

Project AI automatically has access to your entire model structure:

• Every term with its label, unit, value, and metadata
• All blocks and their relationships
• Formula logic and dependencies between calculations
• Cases and how they vary across scenarios
• Financial cascades in the Proforma

There’s no setup, file upload, or data export required. The agent reads your live model and updates as you make changes.

Natural Language Interface

Ask questions through a chat interface directly in the platform, the same way you’d ask a colleague:

“What drives my NPV most?”

“Why is Year 5 cashflow negative?”

“How does the high case for interest rate affect IRR?”

“What happens if I increase system capacity by 20%?”

The agent interprets your intent, identifies relevant terms and blocks, performs any necessary calculations, and responds with specific answers.

Maintains Context

Project AI remembers your conversation within a chat thread, so you can ask follow-up questions naturally without repeating information. It also knows what scenario you’re viewing, which cases are active, and what part of the model you’re focused on.

Examples

Quick Data Lookup

You: “What’s our base case NPV?”

Project AI: “The Net Present Value in your base scenario is $2,547,832.”

Impact Analysis

You: “If we increased the interest rate from 5% to 6%, how would that affect IRR?”

Project AI: “Increasing the interest rate from 5% to 6% would decrease the IRR from 12.5% to approximately 11.2%—a reduction of 1.3 percentage points. The higher discount rate reduces the present value of future cash flows, particularly in later years when most of your project revenue is generated.”

Finding Relationships

You: “What inputs drive NPV most?”

Project AI: “Based on your model, the top 3 inputs affecting NPV are:

1. Electricity Price ($0.12/kWh) — ±15$850K NPV change
2. Construction Cost ($1M) — ±10$320K NPV change
3. Annual Production (20M kWh) — ±10% creates ±$280K NPV change”

Troubleshooting

You: “Why is my Year 3 cash flow negative?”

Project AI: “Year 3 cash flow is negative (-$125,000) because:

• Operating revenue: $480,000
• Debt service payment: $450,000
• Maintenance costs: $155,000

The debt payment combined with maintenance exceeds revenue in this year. Cash flow becomes positive in Year 4 when a major equipment payment completes.”

Building on Context

You: “Show me all revenue assumptions”

Project AI: “Here are your revenue assumptions:

• Electricity Price: $0.12/kWh
• Annual Production: 20,000,000 kWh
• Degradation Rate: 0.5% per year”

You: “What if production was 10% higher?”

Project AI: “If annual production increased to 22,000,000 kWh, your annual revenue would increase from $2,400,000 to$2,640,000—adding approximately $240,000 per year across the project life.”

What You Can Ask

Data and Calculations

• “What’s the current NPV?”
• “Show me all revenue assumptions”
• “How is the payback period calculated?”
• “Find all terms related to solar panel costs”
• “What assumptions are we making about inflation?”

Relationships and Dependencies

• “What drives my NPV most?”
• “Show me all terms that depend on electricity price”
• “Which inputs have the biggest impact on IRR?”
• “Explain the relationship between construction cost and NPV”

Scenarios and Comparisons

• “Compare our base and high cases”
• “What if production was 10% higher?”
• “How does the optimistic scenario differ from pessimistic?”
• “If we increased the interest rate to 6%, how would IRR change?”

Troubleshooting and Analysis

• “Why is Year 5 cashflow negative?”
• “Why is my LCOE higher than expected?”
• “Which assumptions should I validate?”
• “Are there any circular references in my model?”

Limitations

Project AI doesn’t:

• Execute model changes — It can suggest edits but won’t modify terms, formulas, or structure without your action
• Access external data — It works only with information in your model, not live market data or outside sources
• Replace domain expertise — It analyzes your model but doesn’t provide engineering advice, regulatory guidance, or investment recommendations
• Guarantee completeness — Analysis is based on the model as defined; it can’t identify components it doesn’t know should exist

For decisions with major financial or legal implications, verify critical outputs independently.

Privacy and Security

• Project-specific: Each conversation is tied to a specific project—the agent can only see data you have access to
• Organization isolation: Your organization’s data is never shared with other organizations
• Authenticated access: Only users with proper permissions can use Project AI
• Private conversations: Chats are stored within your workspace and not shared across projects or users

Getting Started

Project AI is available in every Aire workspace:

1. Open Project AI — Click the AI icon in the toolbar or navigation menu
2. Start a conversation — Click “New Chat” to begin a fresh thread
3. Ask your question — Type naturally and press enter
4. Build on responses — Ask follow-up questions to dig deeper

The agent automatically indexes your model and updates as you make changes. There’s no setup or configuration required.

Tip: For complex analyses, start with broad questions to establish context, then drill into specifics as the conversation develops.