Writing the memo was never the bottleneck. Defending it is.
An investment committee does not read an IC memo the way a reader reads an article. It goes to the assumptions, asks where the NOI came from, why that cap rate, which comparables support the exit — and the memo's value is entirely in how quickly those questions can be answered.
That is the axis to evaluate these tools on, and it is the one their own marketing pages tend to skip.
The Question A Memo Has To Survive
"Where did this number come from?"
For every figure in the memo, there is a correct answer to that question, and it is specific: a cell in the rent roll, a line in the T12, a named comparable transaction, a field in the public record, or an assumption somebody made on purpose.
A tool that can produce that answer for any figure on the page changes the meeting. A tool that cannot has moved the work rather than removed it — because the analyst now has to reconstruct the provenance of a document they did not assemble.
What To Test, In Order
- Click a number. Take the NOI in the generated memo and ask the tool where it came from. You want the document, the page and the line — not a restatement of the calculation.
- Break an input. Change an assumption and regenerate. Does every dependent figure move consistently, or does the memo now contradict itself in a paragraph nobody reread?
- Starve it. Give it a property with thin comparable evidence. The right behaviour is a stated range and an explanation of why it is wide. The wrong behaviour is a confident point estimate.
- Check the comparables. Are they named and lookupable? Is any of them a listing that never traded, or the subject property appearing under a different address?
- Read the risks section. Generated risk sections tend toward generic language. A useful one names what would have to be true for this specific deal to fail.
The Error That Costs The Most
It is not the number that is obviously wrong. A figure that is absurd gets caught in the first read.
The expensive error is the plausible one: a market rent carried from a neighbouring submarket, an expense ratio that is reasonable for the asset class and wrong for this building, a comparable that is the subject property appearing under a slightly different address. Each of these survives review precisely because it looks right, and each changes the conclusion.
Generation makes this failure cheaper to produce and no easier to catch — unless the figure carries its source. That is the whole argument for auditability, and it is why it belongs above writing quality on any evaluation list.
Why Auditability Is The Differentiator Right Now
The tools in this category — agent builders configured for CRE, memo generators inside deal-management platforms, and general document AI — are converging on similar output quality. Prose is a solved problem.
None of them, on their own published material, demonstrate a figure-level trace back to the source document. That is the gap, and it is not cosmetic: it is the difference between a memo that shortens committee and one that adds a round of verification before it.
Where Titleman Sits
We build the memo from the property record rather than from a prompt. The parcel, ownership and zoning come from public records; the comparables are named with their source; the figures read out of the rent roll and the financials keep a reference to where they were read.
The result is a draft with its provenance attached, for a named professional to review and sign. Where the evidence does not support a number, it says so instead of producing one.
What It Does Not Do
- It does not replace the committee's judgement, and it is not intended to.
- It does not replace institutional lease-by-lease cash flow modelling where the deal requires one.
- It does not replace a licensed appraisal where a lender or a standard requires one.
- A draft is a draft: the professional who signs it is accountable for it, and a tool that obscures which figures are estimates makes that accountability harder rather than easier.