Lease Agreement Analysis: Comparing Terms Across Properties
How AI agents read lease agreements locally and extract comparable terms across a property portfolio for structured comparison and analysis.
The portfolio comparison challenge
A commercial real estate firm managing 40 retail leases across a regional portfolio faces a recurring problem: no two leases are the same, and the differences matter.
Lease A has a 3% annual rent escalation. Lease B ties escalation to CPI with a 5% cap. Lease C has a flat rate for the first three years, then jumps 15% in year four. Lease D has a base rent plus percentage rent based on gross sales.
Multiply these variations across 40 leases, add in differences in CAM charges, renewal options, tenant improvement allowances, exclusivity clauses, co-tenancy provisions, and termination rights, and you have a comparison problem that defies spreadsheets built by hand.
Property managers and real estate attorneys typically track lease terms in spreadsheets that were populated manually when each lease was signed. These abstracts are only as accurate as the person who read the lease, and they degrade over time as amendments, renewals, and side letters modify the original terms without updating the master spreadsheet.
The result: when it's time to evaluate the portfolio -- whether for refinancing, disposition, or simple operational planning -- the team discovers that their lease abstract database doesn't match the actual executed documents. Someone needs to go back to the source and re-read every lease.
Why lease comparison is harder than it looks
Lease agreements resist standardized extraction for several reasons:
No standard structure. Unlike many commercial contracts that follow broadly similar outlines, lease agreements vary dramatically in structure. A lease from a national REIT follows one format. A lease drafted by a local landlord's attorney follows another. A lease inherited from an acquisition follows a third. The same economic terms appear under different headings, in different sections, and with different defined terms.
Layered modifications. A lease that was signed in 2015, amended in 2017 and 2019, and renewed in 2022 has four documents that need to be read together. The current terms aren't in any single document -- they're spread across the original lease and its modifications, with later documents superseding earlier ones where they conflict.
Embedded calculations. Rent escalation formulas, CAM reconciliation mechanisms, and percentage rent calculations are expressed in prose, not numbers. "Base rent shall increase annually by the greater of three percent (3%) or the percentage increase in the Consumer Price Index for All Urban Consumers (CPI-U) published by the Bureau of Labor Statistics for the twelve-month period ending September 30 of the immediately preceding calendar year, provided that in no event shall the annual increase exceed five percent (5%)" is a sentence, not a formula. Extracting the structured financial terms from this language requires reading comprehension, not pattern matching.
Conditional provisions. Many lease terms are conditional. A renewal option might only be available if the tenant is not in default. A termination right might require 12 months' notice and payment of an early termination fee. An exclusivity clause might have exceptions for certain categories of competitors. The conditions are as important as the provisions themselves.
How docrew processes lease portfolios
docrew reads lease documents directly from your file system. For a portfolio analysis, you organize the leases in a folder structure -- one folder per property, containing the original lease and all amendments -- and point the agent at the parent directory.
The agent processes each lease by reading the document in its entirety, understanding the structure, and extracting the terms you specify. Here's the workflow for a 40-lease portfolio analysis.
Step 1: Document inventory. The agent scans the folder structure and identifies all documents. For each property, it catalogs the original lease and any amendments, noting dates and parties. This gives you an immediate view of which properties have clean, single-document leases and which have complex modification histories.
Step 2: Amendment integration. For leases with amendments, the agent reads the original lease and each amendment in chronological order, building a composite understanding of the current terms. When an amendment modifies the rent escalation formula, the agent tracks the supersession. When a renewal letter extends the term, the agent updates the expiration date. The result is an accurate picture of current terms, not a snapshot of the original deal.
Step 3: Term extraction. The agent extracts the specific terms you need for comparison. For a retail portfolio, the typical extraction set includes:
- Base rent (current and historical)
- Rent escalation mechanism and schedule
- CAM charges (fixed, variable, or capped)
- Real estate tax obligations
- Insurance requirements
- Lease term and expiration date
- Renewal options (number, duration, notice requirements, rent terms)
- Tenant improvement allowances (original and any additional)
- Exclusivity provisions
- Co-tenancy requirements
- Percentage rent thresholds and rates
- Termination rights (tenant and landlord)
- Assignment and subletting provisions
- Operating hours requirements
- Maintenance and repair obligations
Step 4: Structured output. The agent compiles the extracted terms into a structured format -- a spreadsheet or structured text file -- with one row per property and one column per term. This is the comparable dataset that the portfolio team needs for analysis.
Extracting financial terms accurately
Financial term extraction from leases is where manual abstraction fails most often. The numbers are buried in dense legal prose, and the formulas are expressed as sentences rather than equations.
docrew reads the financial provisions as a human would -- understanding the language, not just scanning for dollar signs. Consider this common rent escalation clause:
"Commencing on the first anniversary of the Rent Commencement Date, and on each anniversary thereafter during the initial Term, the annual Base Rent shall be increased by multiplying the Base Rent payable during the immediately preceding Lease Year by one hundred three percent (103%)."
The agent extracts: escalation type = fixed percentage, rate = 3%, frequency = annual, compounding = yes, base = prior year rent, start = first anniversary.
Compare that with: "Base Rent for each Lease Year following the first Lease Year shall be adjusted to reflect changes in the Consumer Price Index. The adjusted Base Rent shall be calculated by multiplying the initial Base Rent by a fraction, the numerator of which is the CPI published for the month that is three months prior to the commencement of the applicable Lease Year, and the denominator of which is the CPI published for the month that is three months prior to the Rent Commencement Date."
The agent extracts: escalation type = CPI-indexed, base = initial rent (not prior year), CPI reference = 3 months prior, compounding = no (indexed to initial rent).
The difference between these two formulas is significant over a 10-year term, and it's exactly the kind of detail that gets lost in manual abstraction. When someone populates a spreadsheet with "3% annual increase" for the first lease and "CPI adjustment" for the second, the nuance of compounding versus indexing disappears.
Comparing renewal and termination provisions
Renewal and termination terms are where lease comparison reveals the most value. These provisions determine optionality -- the portfolio's flexibility to adapt to market conditions.
For each lease in the portfolio, docrew extracts the complete renewal and termination picture:
Renewal options. Number of renewal periods, duration of each, notice deadline (critical -- a missed notice deadline means a lost option), rent terms during renewal (fixed increase, fair market value, formula), and any conditions on exercise (no default, minimum sales threshold, landlord approval).
Early termination rights. Whether the tenant has a termination right, the notice period, any termination fee or penalty, and conditions on exercise. A termination right that requires 18 months' notice and payment of unamortized tenant improvements is economically different from one that requires 6 months' notice and no penalty.
Landlord termination rights. Demolition clauses, redevelopment clauses, and other landlord-side termination provisions that create risk for the tenant. These are often overlooked in manual abstraction but can have significant portfolio impact.
The comparison output shows these provisions side by side across the portfolio. This lets the team quickly identify which leases are approaching renewal option deadlines, which have favorable termination flexibility, and which lock the tenant in with limited exit options.
Identifying favorable versus unfavorable terms
With 40 leases abstracted and structured, the portfolio team can now identify outliers -- leases with terms that are significantly more or less favorable than the portfolio average.
docrew supports this analysis by flagging terms that deviate from benchmarks you provide. If your standard CAM cap is $12 per square foot, the agent flags every lease where the CAM provision exceeds that benchmark or has no cap at all. If your standard renewal notice period is 9 months, it flags leases requiring 12 or 18 months' notice.
Common outlier categories in retail portfolios include:
Above-market rent escalations. Leases with escalation rates significantly above current market conditions represent potential renegotiation targets or, if the tenant has a termination right, flight risk.
Below-market rents with long remaining terms. These are portfolio assets -- leases where the tenant is paying below market and has limited ability to terminate. The comparison highlights these as properties where value is locked in.
Missing or weak exclusivity provisions. In retail, exclusivity clauses protect tenants from direct competition within the same property or shopping center. A lease without an exclusivity clause, or one with narrow exclusivity that doesn't cover adjacent categories, is weaker than the portfolio standard.
Unrestricted assignment rights. Most retail leases restrict assignment to maintain tenant mix. A lease that allows assignment without landlord consent could result in an undesirable tenant replacing a desirable one.
Uncapped CAM charges. Common area maintenance charges without caps expose the tenant to uncontrolled cost increases. In a portfolio analysis, these leases stand out as carrying higher cost risk.
The agent presents these findings as a deviation report: for each lease, which terms fall outside the specified benchmarks, and in which direction. This transforms the abstract spreadsheet into an actionable analysis.
Handling different lease formats
The 40 leases in a real portfolio will arrive in multiple formats. Some will be clean Word documents from recent transactions. Others will be PDFs of executed agreements, complete with signature pages and notary stamps. Older leases might be scanned documents where the text is embedded as images rather than searchable characters.
docrew handles format diversity without requiring pre-processing. Word documents are read through the agent's native .docx parser. PDFs are processed through its file reading tools. The agent works through each document regardless of format, extracting the same structured terms.
For a portfolio that includes leases from multiple decades -- common in commercial real estate -- this format flexibility eliminates the document preparation bottleneck. The team doesn't need to convert scans to searchable PDFs or reformat old documents before analysis can begin.
Practical scenario: 40-lease portfolio review
A property management firm is refinancing its retail portfolio and needs current lease abstracts for the lender's due diligence package. The existing abstracts are three years old and don't reflect recent amendments and renewals.
Day 1. The team assembles all lease documents -- original leases, amendments, renewal letters, and side agreements -- into a folder structure organized by property. Total: 40 properties, approximately 120 documents.
Day 1-2. docrew processes the document set. For each property, it reads the original lease and all modifications, builds a composite view of current terms, and extracts the lender's required data points: tenant name, premises, square footage, lease term, rent schedule, escalation mechanism, CAM structure, renewal options, and termination provisions.
Day 2-3. The agent produces a portfolio summary spreadsheet with all 40 leases compared side by side. It flags leases expiring within 24 months, leases with upcoming renewal option deadlines, and any terms that deviate from the portfolio's standard provisions.
Day 3-4. The real estate team reviews the output, validates key terms against the source documents for a sample of leases, and identifies the items that need attorney review -- a lease with ambiguous assignment language, two leases with co-tenancy provisions that may be triggered by a recent vacancy, and one lease where the amendment supersession is unclear.
Without AI assistance, this process takes a paralegal team 3-4 weeks of full-time document review. With docrew, the extraction and comparison work is compressed into days, and the human review is focused on the handful of issues that require professional judgment.
Beyond abstraction: ongoing portfolio management
The initial comparison is the starting point, not the end. With structured lease data extracted from source documents, the portfolio team has a foundation for ongoing management.
Renewal option deadlines become calendar entries with enough lead time to evaluate the market and make informed exercise decisions. Expiring leases become renegotiation opportunities with full visibility into the current terms and how they compare to the portfolio average. Non-standard provisions become action items for the next lease modification.
The key shift is from working with stale abstracts that may not reflect current terms to working with data extracted directly from the executed documents. When someone asks "what are our CAM obligations across the portfolio?" the answer comes from the leases themselves, not from a spreadsheet that was last updated by someone who may no longer work at the firm.
All of this happens on the team's own machines. Lease agreements contain tenant financial data, landlord financial terms, and strategic portfolio information that has no business on a third-party server. docrew keeps every document local, producing analysis that's accurate, current, and private.