Automated close checks find problems faster, but they don’t resolve them. A controlled close depends on classifying each flag, assigning an owner, requiring evidence, and defining who approves and confirms corrections.
Month-end has always had a detection problem. Somebody had to run the reports, scan them, and notice the thing that didn’t look right. Usually that somebody was a senior accountant working late on the third business day.
Yardi just took a big step toward changing that. On October 5, the company announced Virtuoso Enterprise, and one of its agents goes straight at the close. The Month-End Close Agent automatically runs 30 reports across the books, AR, AP, bank reconciliations, security deposits, rent roll, and GL hygiene. It then surfaces issues in plain language with suggested next steps.
For accounting teams, that’s a meaningful shift. The hours once spent hunting for problems can now go toward solving them. And that’s exactly where the opportunity lies. When issues surface faster and more consistently, the value comes from having a clear path for what happens next: who picks up each item, what evidence closes it, and who signs off. Teams that define that path before the pilot starts will get the most out of the agent from day one.
Here’s how we’d approach it.
Every accounting team has a close checklist, even if part of it lives in one controller’s head. Before turning on automated checks, map each one to that checklist. Which existing task does it replace? Which does it supplement? Which does it introduce for the first time?
This matters more than it sounds. If the agent’s bank reconciliation check overlaps with one your staff accountant already performs, you now have two sources reporting on the same account. When they disagree, someone needs to know which one governs. And if the agent flags something no one was formally responsible for, like security deposit liabilities that drifted over several years, you’ve found a gap in responsibility. Not just a gap in the data.
Mapping also shows where checks land across internal staff and outsourced partners. Many owners and operators split the close. Property-level AR might sit with site teams, AP with a shared services group, reconciliations with an outside accounting firm. An automated check doesn’t know where those lines are. Your process has to.
Automated checks are good at finding variance. They’re less good at knowing why it exists. In practice, flagged items tend to fall into three buckets.
Genuine exceptions are real errors: a misposted payment, a duplicate invoice, a deposit recorded to the wrong property. They need correction.
Timing differences are expected: deposits in transit, accruals that will reverse, an invoice received after cutoff. They need documentation, not correction.
Configuration problems are the sneaky ones. A charge code mapped to the wrong GL account. A property set up inconsistently during conversion. A recurring charge that was never ended. These generate the same flag every month until someone fixes the setup. Correcting the transaction each time just treats the symptom.
Teams that don’t classify flags end up doing one of two things. They chase timing differences that would have resolved on their own, or they learn to ignore the same alerts. The second is worse. Alert fatigue is how real errors hide in plain sight.
Once an item is classified, it needs to go somewhere specific. Not to “accounting.” To a person or a role.
For each category of check, define three things up front. Who owns resolution. What evidence closes the item, whether that’s a bank statement, an approved journal entry, a vendor confirmation, or a ticket for a setup change. And when it must be resolved relative to the close calendar.
Evidence is the piece most teams skip. A comment that says “fixed” is not a control. If an auditor, an owner, or a board treasurer asks why a variance disappeared, the answer should be attached to the item, not reconstructed from memory three months later.
Configuration issues deserve their own route. They often need someone with system administration rights, not a staff accountant. If your Yardi environment is supported by an internal systems team or an outside partner, define the handoff between accounting and that team before the first flag shows up.
This is where outsourced and hybrid models get tested. Who can approve a correcting journal entry? Is that authority the same for a $40 rounding difference and a $40,000 misallocation? Who confirms an issue is actually resolved, and can that be the same person who made the fix?
Segregation of duties doesn’t disappear because a machine found the problem. If anything, it matters more, because automation will increase the volume of corrections moving through the system. Set approval thresholds. On material items, separate the person who corrects from the person who confirms. Write down what “closed” means for each category.
Yardi says Virtuoso Enterprise carries over security and governance controls from existing configurations. That’s convenient, but it cuts both ways. If your security groups have grown messy over the years, the agent inherits the mess. Confirm that the people who can post corrections are the people who should be posting them.
Don’t roll out across the whole portfolio in month one. Pick a manageable set of properties or entities, run a few closes, and track a small number of things.
Unresolved exceptions at close. How many flagged items are still open when the books lock, and how old are they?
Repeat issues. Which flags come back every month? Those almost always point to a configuration or process problem worth fixing at the source.
Close time. Not just how fast the checks ran, but how long it took to get from the first flag to signed-off financials.
The goal isn’t zero flags. It’s a close where every flag has a known path to resolution, and fewer of them return next month.
The Month-End Close Agent removes some of the most tedious work in real estate accounting. That’s good news. It frees experienced people for the part that actually requires judgment: deciding what a variance means, whether it matters, and who should act.
But that only works if the operating model is ready. Automation surfaces problems faster. It doesn’t decide who solves them. Teams that settle ownership, evidence, and approval before the pilot will get a cleaner, faster close. Teams that don’t will get a very well-organized list of things nobody fixed