Pursuit · the underwriting pipeline

Reads the deal.Sources every number.Flags what it can't.

A CRE deal goes in as documents and comes out committee-ready in final form based on your standards. Every figure cited to a source. Every gap flagged red for your analyst, never papered over.

Watch a deal get underwritten Start free trial Log in

every number cited or flagged every projection linked to source flags what it can't

The headline walkthrough

2:41 · FULL UNDERWRITE

Watch the underwrite. Not the summary — the build.

One deal, documents to decision. The eight-document set that normally eats the first two days goes in. What comes back opens with its own build sequence — what was read, the rent-roll inputs, net operating income assembled line by line with the basis for each line, the case drivers, five sizing tests, the binding constraint — then the Investment Committee memorandum off the same documents. Every figure expands to the sentence it was read from. Fictional sample deal, labeled as such.

DEEP WALKTHROUGH · fictional sample

Open it in the theater view Start free trial Short on time? The 80-second cut

Product tour

~2 MIN

See the deal room.

A calm walkthrough of the actual product — sign-in, the pipeline staged from screening through closing, a deal workspace with its documents, diligence tasks, key dates and parties, a skill run, and the cited deliverable that comes out of it.

PRODUCT TOUR · fictional sample data

Browse the skill library · 21 recordings → Walkthroughs

Stage 01

Drop the deal.

OM, rent roll, T-12, leases. Every format read, every document typed and parsed. Anything missing gets flagged at the door.

$6.25M net-lease · 4410 S Lamar Blvd, Austin TX · single-tenant CVS · FICTIONAL LIVE RUN

pipeline

auto · hover to pause

The difference

Not a tool. A teammate.

Generic AI waits for you to drive every step and does not know your standards. Pursuit runs the work to your standards and flags what it cannot source.

GENERIC AI CHAT

Any chatbot, driven by hand

You drive every step

Paste, prompt, re-prompt, copy out, reformat. The tool waits for you. The work is yours.

PURSUIT

Your underwriting desk, run for you

The pipeline runs the work

Drop the documents. Extraction, research, model, and memo run end to end. Your analyst reviews the output.

The standard

Every number traces to a source. If it can't find it, it flags it.

Every number is cited to its source or flagged for your analyst. Designed never to fabricate. Every market projection is highlighted in the model and linked to its source, and anything unsourceable is flagged in red, never filled.

Pursuit is designed never to fabricate a number to fill a gap. If a field can't be sourced from your documents or public data, it gets a red flag, and your analyst decides.

◎ source

Source-traced output

Every figure carries a page citation and a verbatim snippet from your own documents.

△ flag

Gap flags, not guesses

Unsourceable values are flagged in red. Never filled in with fabricated data.

◇ audit

Audit trail

Every agent run is logged with the acting user and a timestamp. Reviewable.

fabricated numbers by design

every figure cited to a source page

numbers invented when a source is missing — flagged instead

Sample workflows

The mechanical hours, compressed.

Sample workflows, shown for demonstration. "Before" is the manual process. "After" is Pursuit doing the mechanical work with your analyst reviewing the output. Examples — not client results.

Underwriting from OM see the sample →6-8 hrs → ~30 min

LOI pipeline summary see the sample →half-day → ~10 min

Lease comp resolution see the sample →overnight → same night

Lease expiration tracking see the sample →3 hrs → ~15 min

BOV prep see the sample →4-8 hrs → ~30 min

Off-market owner research · owner LLC → decision-maker + contact path see the sample →manual digging → cited memo

Your workflow herescoped in your first week

Example compressions for demonstration based on the sample pipeline. Every output still ends in analyst review.

Ready to start