How does Vic.ai make predictions?
Vic.ai invoice predictions are made by artificial-intelligence (AI) models based on machine learning (ML).
Using these AI models, everything is predicted including vendors, total amounts, invoice numbers, invoice dates, due dates, and payment terms at the invoice level.
At the line item level we predict invoice items, amounts, GL accounts, and dimensions.
What is Autopilot?
Our autonomous accounting service, Autopilot, is enabled on a per-company basis once the AI models have processed enough data and are confident in their predictions. Invoices on Autopilot have an accuracy rate of 95% or higher.
How do I get to Autopilot more quickly?
Machine Learning-based AI learns continuously from your input. This means that all corrections will instruct the AI on how better predict in the future. This feedback will increase accuracy over time, as the AI balances between optimally learning from the latest feedback and the millions of invoices it has seen previously. The more invoices the AI sees from your company, the higher confidence of automation you will get, and the fewer corrections will be required.
What prevents/slows my client from launching Autopilot?
The #1 reason that Autopilot takes longer to launch is that predictions are changed with arbitrary edits. Arbitrary edits are changing values to those not found on an invoice. Read the full Arbitrary Edit article here!
How do I Autopilot more of my client's invoices?
Invoice volume, consistency, and not making arbitrary changes.
Invoice Volume: More data points allows the AI to be more confident in its predictions.
Consistency: Coding invoices the same way each time.
Arbitrary Changes: Even if the client requests invoices in a certain manner, or to be paid on a cash.
Looking for input on how to get your company on Autopilot quicker? Contact Vic.ai Customer Support for tips and tricks on getting your company in the processing fast lane!