
Rivian’s AI Agents Cut 15 Days From Every Close Cycle
Rivian has automated a key finance task tied to custom vehicle tooling, cutting more than 15 days of manual work per close cycle. The system tracks purchase order accruals for tools ordered but not yet billed, using AI on Amazon Bedrock and AWS.
What the AI automates in finance close
Custom manufacturing tools for EV production, such as stamping dies and injection molds, can take one to two years to build. Rivian often does not receive invoices until 18 months or more after placing orders, while accounting requires recognizing costs gradually over time.
Rivian’s AI system monitors purchase orders needing accruals, calculates what is owed, and drafts accounting entries for approval.
How Rivian avoided hard-coded accounting rules
Instead of hard-coding fixed automation logic, Rivian stored its actual accounting procedures as plain-written instructions that the AI reads directly. The approach was designed to handle rule changes that vary by purchase size and type.
A finance manager can update the process by editing the written instructions, without rewriting code. The AI applies the updated instructions automatically in subsequent runs.
Human sign-off and feedback loop for errors
A human still approves every entry before it posts to Rivian’s books. Rivian also built in a feedback mechanism: when a manager corrects an error, the correction feeds back into the instructions so the same mistake should not recur.
Rivian said auditors found the new process easier to trace than the spreadsheets it replaced.
Built in weeks, expanded for more finance work
Rivian built a proof of concept in five weeks before moving to real use, according to an Aug. 13 Amazon Web Services (AWS) blog post. The company is already planning to apply the same approach to other finance tasks.
Why it matters
By shifting accrual tracking from manual spreadsheet work to AI-driven, instruction-based automation, Rivian is reducing close-cycle time while keeping controls and audit traceability intact.