DealWise AI
Product

Upload the documents. The deal builds itself.

Drop in the offering memorandum, the leases and the operating statement. DealWise reads them and proposes the property facts, the suites and the rent schedule — you review an answer instead of assembling one.

What happens

A blank form is the slowest part of underwriting

Most underwriting software starts you at an empty model. The modelling that follows is often excellent, and it begins after the slow part is already done.

  • Drop in what you were sentThe offering memorandum, the executed leases and every amendment stacked on them, the operating statement. Scanned PDFs included.
  • It reads them, then proposesProperty facts, suites, and each suite's own stepped rent schedule with its dates and lease type — proposed and filled, not silently written.
  • You review, edit and commitEvery value is yours to change before anything is saved. Reviewing a filled model is a different job from building an empty one.
The part that matters

A lease is a contract. An offering memorandum is a pitch.

Reading a PDF is not the hard part, and it is not the claim. The claim is that what a document is decides what the app is allowed to do with what it says. Three tiers, and the third is the one that earns its keep.

  • Contract terms — proposed and filledSuite, tenant, square feet, dates, rent and lease type come out of the lease with the sentence they came from and the page it was on.
  • Reported figures — filled, with their sourceOperating-statement lines are proposed the same way and labelled with the statement and the period they cover. An expense you cannot trace is one you cannot defend to a lender.
  • The seller's claims — captured, never writtenAsking price, stated cap rate, pro-forma NOI, market rent, growth. All transcribed, all displayed beside your deal, none of it an input to it.

Why keeping the seller's numbers out is what makes them checkable

It reads like a restriction and it is actually the feature. Because the stated cap rate never becomes an input, it stays an independent claim — and an independent claim can be measured.

That is where the stated versus derived check comes from: the gap, in basis points, between the cap rate on the flyer and the one your rent roll and the real expenses actually support. A model that had quietly absorbed the sellers NOI could not produce that number, because it would be comparing the claim against itself.

The same logic runs through the rest of the product. NOI is always derived from the rent roll and never entered beside it, which is why a figure here reconciles to an operating statement instead of merely looking plausible.

Questions people ask

What documents can I upload?

The offering memorandum, executed leases and their amendments, and the operating statement — the trailing twelve months or a year-end. Rent rolls and comp sheets too. PDFs are the normal case, including scanned ones.

What does it fill in automatically?

Property facts, the suites, and each suite's rent schedule with its dates, steps and lease type. On the expense side, the income and operating expense lines off the statement, with capital items separated from operating ones. You review everything before it is committed.

Does it use the cap rate printed in the offering memorandum?

No, and that is deliberate. The asking price, stated cap rate, pro-forma NOI, market rent and growth assumptions are captured and shown beside your deal, but none of them becomes an input to it. Your NOI comes from the rent roll and the real expenses. Keeping the seller's claims out of the model is exactly what makes it possible to measure them against it.

How do I know the extraction is right?

Every extracted term is shown against the clause it came from, with the page it was on. Checking one takes a glance rather than a re-read. Extraction is AI and it is not perfect — the answer to that is that you can see the source for every value, not that we claim you will never need to look.

What happens with a scanned lease that has no text layer?

It is read as an image. Long documents are split into passages and the extraction runs across all of them, so a rent step buried on page fifty-one of a third amendment is found the same way one on page two is.

Is this faster than just typing it in?

That is the whole point of it. Getting terms out of paper is where underwriting time actually goes — the modelling is quick once the rent roll is right. Starting from a document rather than a blank form is what makes a first pass cheap enough to run on every deal instead of only the ones that already look good.

Start with a document, not a blank form.

Upload an offering memorandum and a stack of leases. DealWise reads them, builds the rent roll, models every month of the hold and ends with a buy, hold or pass — with the clause it relied on. Free plan, no credit card.

Upload & Auto-Build a CRE Deal | DealWise AI