Finding Candidate Clauses Across a Large Lease Set

CRE teams often need to locate leases that may contain co-tenancy triggers, radius restrictions, exclusive-use provisions, or similar clauses. A targeted AI query can help surface candidates from retrieved evidence, but it may not inspect every lease, page, or amendment. Confirm the intended review set and verify each material finding.

The Search Problem

Commercial leases do not use standardized language. One lease may call something a “co-tenancy provision,” another a “co-occupancy requirement,” and a third may describe the concept without a heading. Semantic retrieval can broaden the candidate set beyond exact keywords, but it does not guarantee exhaustive coverage.

Why These Clauses Hide

A portfolio's lease set is not one document repeated; it is decades of negotiation by different counsel, on different forms, under different market conditions. That history is exactly what makes the search hard:

• The operative language often lives outside the base lease — in a second amendment, a side letter, or an estoppel — while the base lease reads clean.
• A provision can be drafted as its remedy rather than its trigger: a co-tenancy right may appear only as a rent-reduction formula, with the word “co-tenancy” nowhere on the page.
• Defined terms send the reader elsewhere. A radius restriction measured from a point defined in an exhibit means the clause and its real content sit pages apart.
• Older leases exist as scans of varying quality, so the text a search engine sees may be an imperfect rendering of what the parties signed.

Each of these is a reason a keyword search misses, and each is also a reason a semantic candidate list still needs a reader who knows what the clause does, not just what it is called.

Questions to Investigate

Using Results in Due Diligence

Hidden lease provisions can create material acquisition risk. AI can surface candidate clauses with source evidence for review; it can also miss relevant documents or misinterpret language. Validate coverage, read the governing provisions, and obtain qualified review before relying on a result.

The discipline that makes candidate lists useful is the same one that governs any document review: know the denominator. Before treating a result as a survey of the portfolio, confirm which leases — and which amendments to those leases — were actually in the indexed set, and reconcile that list against the deal's own document inventory. A candidate list is a starting inventory to be disproven by reading, not a certification that nothing else exists; the leases where the query returned nothing deserve a reader's attention in proportion to what is at stake. Recording what was checked, by whom, and against which document set turns a fast search into something a diligence file can actually rely on.