A property-management back office that re-keyed every owner's paperwork by hand
Built one client-file platform as the single source of truth, with an ingestion-and-classification step that reads incoming owner and vendor documents however they arrive and files them against the right property, plus a reconciliation pass that flags mismatches and duplicates instead of leaving them to a person to catch.
The firm manages a large book of properties, and every owner sends paperwork their own way: some email a scan, some drop files in a portal, a few still hand over paper. Statements, invoices, lease changes, vendor bills, all of it arrives in a different shape, and the back office reads each one, decides which property it belongs to, and types it into the systems by hand. The volume, not the difficulty, was the problem: the same numbers got entered more than once, in more than one place.
The tools made it worse. New work was tracked in one system, then the same client and property details were re-keyed into a second system used as the records database. Earlier attempts at automation had stalled at single-purpose scripts, an OCR step here, a CSV export there, none of it joined up, so the team had hit the ceiling of what they could stitch together on their own.
We started with the foundation the team asked for: one platform holding every property and every owner as a single client file, so the same fact is entered once, at the right place, and everyone is looking at the same record. Structured intake, the parties on each file, role-based access, and a dashboard showing the status of every file at a glance.
On top of it sits the document step. An incoming statement, invoice, or vendor bill is read whatever the format it came in, classified, and filed against the right property with its key fields pulled out, so the back office reviews and confirms rather than retypes. A reconciliation pass then compares what was ingested against what is on file and surfaces mismatches and likely duplicates for a person to judge. The extraction is AI-driven with a deterministic fallback behind it, so a messy scan still produces a usable draft and never simply fails. Sensitive financial data stays inside the firm's own privacy boundary.
The double entry is gone. A document now arrives, gets read and filed against the right property automatically, and the back office spends its time confirming and handling exceptions instead of retyping numbers from one system into another. That gave back roughly two days a week across the team, without adding people.
The records stopped drifting, too. There is one client file per property now, the status of every file is visible in one view, and the reconciliation pass catches the mismatches and duplicates that used to surface late, or not at all. New staff find their way around in days rather than weeks.
“The work was never hard, there was just too much of it, and we were typing the same numbers into two systems. We had automated little pieces ourselves and hit a wall. We needed it to actually join up.”
Have a mess like this?
Tell us where time and money are leaking. We’ll map the fastest high-ROI win, usually a Pilot that pays for itself.
