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

# Lease abstraction

> Run and review AI-assisted lease abstracts with source-backed fields.

Lease abstraction turns a commercial lease (and related amendments) into a structured summary your team can review and use operationally — terms, dates, options, co-tenancy, **CAM** (common area maintenance) language, and other fields your workspace is configured to capture.

**Hugo** is the agent that runs the workflow. If you are new to Surfaice, read [What is Surfaice?](/concepts/what-is-surfaice) and [Skills](/guides/skills) first.

## Before you start

* You need access to a workspace with the lease-abstraction skill enabled.
* Prefer uploading or pointing Hugo at the authoritative PDF your team already files in SharePoint or your document management system.
* Confirm exclusion lists and connector scopes with IT if the packet includes sensitive side letters.

## Typical flow

<Steps>
  <Step title="Open the skill or ask Hugo">
    Start lease abstraction from the skills UI, or ask Hugo to abstract a specific lease document.
  </Step>

  <Step title="Provide the source documents">
    Attach the lease PDF and relevant amendments. Include exhibits only when they affect commercial terms you care about.
  </Step>

  <Step title="Review field-by-field">
    Check dates, rent, options, co-tenancy, and landlord obligations. Prefer fields that cite page or clause references.
  </Step>

  <Step title="Correct and save">
    Fix errors in review. Save only the accepted abstract into the workspace or write back to your lease admin system if that path is enabled for your tenant.
  </Step>
</Steps>

<Warning>
  Treat the first pass as a draft. High accuracy on standard retail leases still requires human review before downstream reporting or landlord correspondence.
</Warning>

## Tips for better abstracts

* Use clean scans or native PDFs when possible. Scanned retail leases are supported, but low-quality scans may miss fields — compare to the signed PDF.
* Include amendments in chronological order.
* Tell Hugo which field set your team uses (for example "JD retail abstract fields") if multiple templates exist.
* Ask follow-ups like "quote the exclusive-use clause with page number" when a field looks incomplete.
* Say **transcribe only** if you need OCR Markdown before abstraction.

## Scanned leases

Many tenants work from image-only PDFs. Hugo can transcribe and abstract them, but:

* Run `ocr-transcription` first when you need searchable text without field extraction
* Expect extra review time on handwritten riders and stamp overlays
* Ask which files were read if a portfolio folder contains hundreds of scans

## Related

* [Agent limitations](/concepts/hugo-limitations)
* [Working with Hugo](/guides/working-with-hugo)
* [Chat with Hugo](/guides/chat-with-hugo)
* [Skills](/guides/skills)
* Marketing deep-dive: [Complete guide to lease abstraction](https://www.surfaice.pro/blog/complete-guide-lease-abstraction)