Skip to main content
SevrelDocumentation
← Back to App
Documentation menu

How Search & Retrieval Works

When you ask Sevrel a question, it doesn't guess the answer. It searches your actual documents, reads the relevant sections, and builds a response grounded in what it found. Here's how each step works.

Document Search

Sevrel uses a technique called Retrieval-Augmented Generation (RAG). Unlike a general-purpose chatbot that relies on pre-trained knowledge, RAG forces the AI to base its answers on your specific documents. This dramatically reduces hallucination and ensures responses are grounded in real data.

1

Document Search

Sevrel searches your Egnyte library for files matching your query. It looks at file names, folder structure, and document content to find the most relevant sources.

2

Text Extraction

The relevant documents are opened and their text content is extracted. Sevrel handles PDFs, Word documents, Excel spreadsheets, and other common formats used in commercial real estate.

3

AI Analysis

The extracted text is sent to the AI model along with your question. The model reads the content and generates an answer based specifically on what the documents say.

4

Source Citation

Every factual claim in the response is linked to the specific document it came from. Click any citation to see the source file name and relevant excerpt.

What Documents Does Sevrel Understand?

Sevrel can read and analyze most document formats used in commercial real estate workflows:

Fully Supported

  • PDF documents (including text-based PDFs)
  • Microsoft Word (.docx)
  • Microsoft Excel (.xlsx) — reads structured data
  • Plain text files

CRE Document Types

  • Lease agreements and amendments
  • Rent rolls and tenant rosters
  • Operating expense budgets and reconciliations
  • Shared cost adjustment reports
  • Property appraisals and inspections
  • LOIs and purchase agreements

How to Get Better Search Results

The quality of Sevrel's answers depends on how well it can find the right documents. Here are strategies to improve results:

  • Name the property. “What is the base rent at Pembroke Lakes Square?” is much more effective than “What is the base rent?”
  • Mention the document type. “Check the 2025 shared cost adjustment for Bal Harbour Square” helps Sevrel narrow its search.
  • Include dates or time periods. Financial data often exists across multiple years. Specifying “Q4 2025” or “2024 budget” reduces ambiguity.
  • Use tenant names. When asking about a specific lease or tenant, include their name exactly as it appears in your documents.

Real-Time Thinking Indicator

While Sevrel processes your query, a live timeline shows exactly what it's doing at each step — searching for documents, reading specific files, and synthesizing an answer. This transparency lets you see that Sevrel is working with real data, not guessing.

Searching for “CAM 2024 Bal Harbour”
Found 8 files — reading CAM Budget 2024 Bal Harbour.pdf
Generating answer from 3 source documents

This also helps with troubleshooting — if Sevrel searched the wrong documents, you can rephrase your query with more specific property or tenant names.

Search-First Efficiency

Sevrel uses a search-first approach rather than scanning every document in your library. It sends targeted search queries to your Egnyte account, reviews the search results for relevance, and only reads the full text of documents that are likely to contain your answer. Most queries complete in 2-5 steps.

This means new files are available immediately — no indexing delay. As soon as a document appears in your Egnyte folders, Sevrel can find and read it.

Understanding Source Citations

Every response from Sevrel includes numbered citations that link back to the source documents. This transparency is critical for CRE work where accuracy is non-negotiable.

Example response:

“The base rent for T-Mobile at Pembroke Lakes Square is $45.00 per square foot 1, with annual escalations of 3% 2.”

Clicking [1] reveals the source: “T-Mobile Lease Amendment 2024.pdf — Section 4.1: Base Rent”

Next Steps

Last updated: March 17, 2026