The real estate back office is still a spreadsheet

Every commercial lease is a set of instructions for collecting money. Rent steps up on a set date. The tenant pays its share of taxes, insurance and common area costs. Some leases cap those charges. Some exclude certain expenses. Some set a base year.

The instructions are clear. The problem is that nobody carries them out as written.

Here is how it usually goes. A lease is signed and filed as a PDF. Someone keys the main terms into the property management system. Base rent goes in. The escalation date might go in. The pass-through clauses, the caps, the exclusions and the base year live in the PDF and in the head of whoever read it. Then the months go by.

At month end, an accountant opens the bank statement, the rent roll, a stack of vendor invoices and a spreadsheet. They tie one to the next by hand. They catch what they can. When the close is done, the owner gets a report. It looks precise. Often it is a careful estimate with clean formatting.

Where the money goes

Money leaks out in the gap between the lease and the ledger. It happens in the same three places, over and over.

Missed escalations. The rent was supposed to step up in March. The system still bills the old amount in July. Nobody notices, because nothing looks broken. The tenant pays what it is billed.

Unbilled pass-throughs. The lease says the tenant pays its share of operating expenses. Reconciling those expenses at year end takes real work. Someone has to pull every invoice, sort what is recoverable and apply the caps and exclusions. When the work is hard, it runs late. When it runs late, it gets skipped or settled.

The manual close. Each month, skilled people spend days matching numbers a machine should match. The cost is not only their time. It is the errors that slip through when tired people do repetitive work, and the decisions owners make on numbers nobody can trace.

Across the industry, Komod’s model puts this leakage at about $14 billion a year. The exact figure matters less than its shape. It is not one big mistake. It is thousands of small ones, every month, in every portfolio.

Why “AI that helps” is not enough

The obvious answer is software that helps. A chatbot that answers lease questions. A tool that summarizes a document. A dashboard that flags anomalies.

I have sold software and I have bought it. Through St Lue I have watched companies pay for AI and see very little back. The tools get bought. The teams keep working the old way. The spend outruns the return.

Helping is the wrong goal. A lease summary does not bill the escalation. A flag does not reconcile the invoice. If a person still has to do the work after the AI is done, you added a step. You did not remove one.

The work itself has to get done. Read the lease. Pull out every term that moves money. Match every invoice and every bank line. Work out what each tenant owes. Prepare the close.

Two rules make it safe

Doing the work only matters if you can trust it. That takes two rules.

First, every number traces to its source. If the close says a tenant owes a certain amount for common area costs, you should be able to click through to the lease clause, the invoices and the math behind it. No black box. An owner, a lender or an auditor should be able to check any line in a minute.

Second, a person approves the money. Software prepares. People decide. Nothing gets billed, paid or booked until someone accountable has looked at it and said yes. Good finance teams already work this way. The software should fit that habit, not route around it.

With those two rules you can hand over the work and keep the control.

What we are building

This is why we built Komod.ai. Komod closes the monthly books for commercial real estate owners. It reads leases, invoices and bank statements. It reconciles every line. It prepares the close for a person to approve, and every number ties to its source document.

Property managers should manage property. They walk the buildings, look after tenants and handle what breaks. They should not spend their week keying invoices into a spreadsheet.

We are onboarding our first paying commercial real estate customer now. There is a lot left to build. The direction is clear.

The back office should not be a spreadsheet. It should be a system that does the work, shows its sources and waits for your approval.