How does AP Autonomy work?

Jessica Sadow

Jessica Sadow

Director of Product Marketing

How does AI work to make invoice processing autonomous? Learn how's AI manages the accounts payable process end-to-end, emulating human decisions.

July 26, 2023

5 min read

How does AP Autonomy work?

What is the definition of AI

When companies talk about AI, what do they really mean? As we’ve explored previously, generative AI is not the same as automation. In textbook terms, artificial intelligence (AI) technologies can be used to streamline the entire invoice management and processing workflow. It leverages advanced machine learning algorithms, optical character recognition (OCR), natural language processing (NLP), and data extraction techniques to extract relevant data from invoices, validate and verify the information, and perform various tasks associated with invoice processing.

But how does this help AP departments?

How AI can solve AP department challenges

By leveraging AI, AP departments can minimize challenges with inefficiency and potential for error in manual data entry, expectation handling, and visibility when it comes to invoice processing that goes beyond simple AP Automation.

  1. Manual Data Entry and Processing: According to a recent study, approximately 25% of the average accounts payable professional's daily time is spent on tasks that could be automated. These tasks include typing invoice data, fixing mistakes, and responding to calls and emails about the status of invoices and payments. This traditional approach to accounting systems consumes valuable time and is susceptible to errors. 
  2. Invoice Exceptions and Discrepancies: Up to 24% of AP teams' time is spent on exception handling. Such exceptions can encompass missing information, incorrect pricing or quantities, disparities between invoices and purchase orders, or non-compliance with established processes. This can result in delayed processing, late payments, and penalties.
  3. Lack of Visibility and Control: PYMNTs reports that 71% of AP professionals raised concerns about a lack of visibility into manual workflows. Without an efficient system in place, tracking the status of invoices, identifying bottlenecks, and ensuring adherence to payment terms become arduous tasks. Insufficient visibility can result in fraud, missed payment deadlines, late fees, and difficulties managing cash flow.

Our focus at is to solve these pain points associated with invoice processing. For example, we’ve spoken to a lot of organizations who say they feel they are not well-equipped to handle an increase in invoice volume. This is where steps in to manage general ledger (GL) coding, PO matching, and more, autonomously, on behalf of the AP team.

How AP Autonomy works’s AI runs the accounts payable process from beginning to end, mimicking human decision-making. Invoices get automatically ingested, classified, and sent to approvers without requiring any data entry or classification review beforehand. Let’s break it down into each step in the process.


The system can ingest invoices through various methods such as email, manual upload from the desktop, snapping a picture using a mobile app, or through EDI, API, or SFTP.

Invoice processing

Once the invoice is ingested into the system, the AI starts processing the invoice. It makes predictions on two aspects of the invoice: the header-level data (like invoice number, due date, terms, amount, currency) and the line-item level data (like GL Account, location, department). If there is a PO associated with the invoice, it will be automatically 2 or 3-way matched.

Human review and approvals

The AP staff reviews the predictions made by the AI and confirms if they are correct or makes edits if necessary. The AI learns from these interactions and improves its accuracy over time. Once the AP staff validates the predictions, the invoice is sent for approval. The approver can view and approve the invoice directly from the mobile app.

Autonomous approvals

Once the AI is extremely confident across all predictions, the invoice can go straight from ingestion into either an approval flow or straight into the ERP ready for payment. The longer clients use, the more autonomy they get. The platform operates at 97-99% AI accuracy, surpassing human-level accuracy.


When using Payments from, customers can manage all payment types seamlessly: check, ACH, and virtual card. Confirmed payments are posted in the ERP system as paid & in the Bill Pay view.

ERP Integrations

Vic integrates with many ERPs out of the box today, including Netsuite, SAP, workday, Microsoft Dynamics, PeopleSoft and more.

Intelligent analytics Intelligence ensures that autonomous invoice processing doesn’t turn into a black box by surfacing insights in real time. It provides customers with customizable processing insight dashboards and includes performance data based on user, region, AI accuracy, processing time, and more

Conclusion's proprietary AI can autonomously learn and adapt to process invoices with high accuracy, reducing the need for constant human supervision.

In fact, the longer clients use, the more autonomy they get. The platform operates at 97-99% AI accuracy, surpassing human-level accuracy, and clients see up to 355% improved invoice processing capacity per full-time employee.

New call-to-action


Sign up to receive the latest news, updates and valuable accounting and finance resources conveniently in your inbox.