Your Own Timecards Are Now the Plaintiff's Exhibit A
On August 18, litigation-support company Anytime AI launched a "Staffing Analysis" module that lets plaintiff firms compare a nursing home's federally reported Payroll-Based Journal (PBJ) staffing data against the facility's own timecard and payroll records, calculating Hours Per Resident Day by nurse category and flagging day-by-day shortfalls. CEO Teddy Wu says the pain point is that firms "know a facility was understaffed, but proving it meant waiting months and spending thousands of dollars on an outside expert." The release names four plaintiff firms as early users: Tosh Law Firm, Moore Hutchins Moore, Henson Fuerst, and Garcia & Coman (GlobeNewswire release carried by the Manila Times, Aug. 18, 2026). Read it for what it is: a vendor-published launch announcement, with self-reported capability claims and vendor-supplied names.
Plaintiff firms have been using AI to mine ownership and clinical records all year, so the trend itself is not new. What is new is that the trend now has a name, a vendor, and a price, and it is pointed at one specific distance: the gap between what you report to Medicare and what your own punch clock says.
Understaffing litigation used to run on general allegations and an expensive expert reconstruction. Months of it. That cost was, functionally, a filter. Now it is a product on a shelf, marketed by name to the people who sue you, and it closes that gap in an afternoon.
Read the sentence one of those attorneys gave the vendor for its own launch release. "It sent me to the exact cell in the PBJ worksheet to back up every finding, and I was able to take that data straight into my demand letter." Sit with the picture for a second. A lawyer at a desk, your quarterly filing open in one window, your punch clock in the other, and the software walking him to the cell. He did not subpoena anything. He did not hire anybody. He read the homework you turned in.
That is exposure twice over. The first layer is the shortfall, if there is one. The second layer is worse, because the weapon was built out of records you were already required to generate and file. It does not land like a new technology risk. It lands like a file cabinet turning against you.
Which is why governance can no longer stop at "how are we using AI." It has to include "how would an adversary use what we are already required to produce." That means a standing PBJ-to-timecard reconciliation, owned at the leadership level, not a posture that waits for a subpoena to check the math.
This month's action: pull your last two quarters of PBJ submissions and reconcile them against actual timecard and payroll data yourself. Where gaps exist, write down the operational reason now, while it is still a management note and not a deposition exhibit.