In development
Example task
Month-end close
Conceptual example
Receive the books
Entries, statements, and support
Reconcile accounts
Match each balance to its support
Close the period
Statements tie out
What the task measures.
Score whether every account reconciles and the statements tie out. Record each entry the model posted and each document it relied on.
Totals can balance while individual entries are wrong. Checking each reconciliation against its support tests whether the model did the work correctly.
Work with Kimpton.
Engagements are custom. We agree on the task, inputs, outcomes, and rewards before we build the environment.