How a global insurance group reconciles 340 accounts across 12 entities overnight
With 12 entities from rapid acquisitions and 340 accounts to reconcile monthly, the close was a two-week scramble. Cortex now handles the investigation overnight.
Five hundred acquisitions built this insurance group into an 18,000-person platform. Each deal added entities, charts of accounts, and balance sheet accounts that needed monthly reconciliation. Six people spent two weeks every month on manual tick-and-tie across 340 accounts. Errors compounded. Auditors flagged reconciliation quality two years running.
Cortex changes the equation
The VP chose Arvexi for Cortex, the AI investigation engine. Unlike rule-based tools that flag exceptions for humans, Cortex performs the investigation: pulling GL detail, comparing against expected balances, identifying root causes, and generating findings with deterministic confidence scores.
"The first month, Cortex completed overnight what used to take six people two weeks," the VP says. "Every account had a status, every variance had an explanation. Our team reviewed results instead of building them."
The account reconciliation module handles matching and status tracking. Cortex handles the judgment-intensive investigation. The deterministic scoring means auditors can trace any finding back to the underlying data and verify the logic independently.
Results
The close dropped from 12 days to 5. Six FTEs were redeployed to FP&A. Unresolved variances at close fell to zero for the first time.
Explore how Arvexi's account reconciliation platform and Cortex AI transform the monthly close for multi-entity organizations.
We reconcile 340 accounts across 12 entities every month. Before Arvexi, that meant 6 people working overtime for two weeks. Now Cortex handles the investigation overnight. Our team reviews findings in the morning. The work is already done.
VP of Finance
Finance Leadership at Global Insurance Group