> ## Documentation Index
> Fetch the complete documentation index at: https://docs.crefi.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Key Concepts

> How CREFI organizes your data and workflows

Understanding a few core concepts will help you get the most out of CREFI.

## The Deal Graph

CREFI organizes everything around **deals**. A deal is the central workspace that ties together properties, stakeholders, financial data, and your analysis.

<AccordionGroup>
  <Accordion title="Deals" icon="handshake">
    A deal represents a potential or active transaction. It progresses through a lifecycle:

    **Draft** → **Underwriting** → **Active** → **Closed** → **Archived**

    Each deal contains properties, entities, uploaded documents, chat conversations, and workflow outputs.
  </Accordion>

  <Accordion title="Properties" icon="building">
    Properties are the real estate assets tied to a deal. Each property has:

    * Address and geographic data
    * Property type (multifamily, office, retail, industrial, etc.)
    * Key metrics (units, square footage, year built)
    * Photos and associated documents

    Properties can be marked as the **subject property** (the one you're analyzing) or as **comparables** (similar properties used for benchmarking).
  </Accordion>

  <Accordion title="Entities" icon="users">
    Entities are the people and organizations involved in a deal:

    * **Buyers** and **Sellers**
    * **Investors** and **Lenders**
    * **Tenants**
    * **Builders** and **Developers**

    Tracking entities helps CREFI understand the full picture of a transaction.
  </Accordion>

  <Accordion title="Transactions" icon="money-bill-transfer">
    Transactions are financial events associated with a deal:

    * Purchases and sales
    * Leases
    * Loans and refinancing

    These give CREFI the financial context it needs for accurate analysis.
  </Accordion>
</AccordionGroup>

## Documents and Files

CREFI processes the documents you upload so the AI can reference them in conversations:

| File Type                             | What CREFI Extracts                      |
| ------------------------------------- | ---------------------------------------- |
| **PDF** (offering memos, appraisals)  | Full text, tables, and key data points   |
| **Excel** (rent rolls, T-12s, models) | Cell data, formulas, sheet structure     |
| **Images** (property photos)          | Associated with properties for reference |

Files go through automatic processing after upload. Once complete, their contents are available to the AI across all conversations in that deal.

## Workflows

Workflows are guided, multi-step processes that help you complete complex tasks. Instead of free-form chatting, workflows walk you through structured inputs and produce specific outputs.

CREFI currently offers two workflows:

<CardGroup cols={2}>
  <Card title="Underwriting" icon="chart-line" href="/features/underwriting">
    Takes your property data, assumptions, and financial documents to produce a populated underwriting model.
  </Card>

  <Card title="Investment Memo" icon="file-lines" href="/features/investment-memos">
    Generates a comprehensive investment memorandum with sections like executive summary, market analysis, and financial analysis.
  </Card>
</CardGroup>

## Assumption Templates

Assumption templates are reusable sets of underwriting assumptions (growth rates, cap rates, expense ratios, hold periods, etc.) that you can apply to deals. This lets you standardize your analysis across your team while still adjusting assumptions per deal.

## Memory

CREFI learns from your conversations. When it encounters valuable insights, preferences, or data points, it can store them in its **memory**. This means:

* The AI remembers context from past conversations
* Analysis quality improves over time as it accumulates knowledge about your deals and preferences
* You can view and manage stored memories from the Memory page