FAQ: Recognition


ElevateAP's Recognition translates the most important info from your invoices into data that can be submitted to Workday. It uses sophisticated logic, machine learning, and AI to identify patterns in your documents and match them to your existing Workday data. Over time, it refines its methods so they become specific to your organization.

Recognition is constantly evolving. ElevateAP has multiple recognition methods, so it chooses the service that is the most successful for each specific field or supplier.

Recognition Rate

While ElevateAP attempts to predict all fields, we prioritize reporting on the following fields:

  • Supplier ID
  • PO Number
  • Invoice Date
  • Invoice ID
  • Company ID
  • Control Amount
  • Freight Amount
  • Tax Amount
  • Currency

All Recognition Rates are calculated by taking the number of priority header fields that were correctly predicted on an invoice and dividing it by the total number of priority fields.

Recognition Rate = Total priority fields correctly predicted / Total priority fields analyzed x 100

The total number of priority fields can vary based on the timeframe and the number of invoices with a valid supplier. Learn more about Analytics in ElevateAP.

Frequently Asked Questions

How can we improve Recognition Rates?

For best results, make sure that you’re following the recognition best practices, especially the suggestions for Workday data and settings. Improving recognition starts with making sure that your Workday data is accurate so that it can be matched to your invoices.

We also have a list of best practices you can share with your suppliers. It includes suggestions for how they can optimize their invoices for Ascend’s Recognition. Your Customer Success Manager can provide expert guidance to help improve recognition rates beyond best practices.

When can we expect Recognition Rates to improve?

Recognition Rates usually improve with increased invoice volume. The more invoices you receive, the faster ElevateAP can learn. The most significant improvements will likely be for invoices from suppliers that use consistent invoice formats and the suppliers that invoice you most frequently. Improvements may be slower for suppliers with more complex or varied invoice formats.

Additionally, you may initially see fluctuations in overall recognition rates. This is normal and expected as ElevateAP refines its methodologies for new types of invoices it hasn’t previously analyzed.

Why don’t certain fields get picked up?

There could be several reasons why a field isn’t being recognized. The invoice image could be in an irregular format, have low resolution, or the file may be corrupted. The field may be one that ElevateAP doesn’t typically pick up. Oftentimes, the field is simply missing from the invoice. Make sure you let your suppliers know if there is a field missing.

Will the system learn to pick up a specific field over time?

It’s possible, but it depends on your invoice volume. Recognition improves as your invoice volume increases. That’s why your most common invoices generally have the best recognition rate—ElevateAP has had the chance to learn from a larger data sample. When ElevateAP is able to access more data to learn your invoices, it can make more accurate predictions.

If you’re experiencing any recognition issues related to a specific type of invoice or supplier, let us know.

Does ElevateAP recognize handwritten text?

No, ElevateAP does not support handwritten text on invoices. See our full list of supported files.

What languages does ElevateAP recognize?

ElevateAP supports the following languages:

  • English
  • German
  • French
  • Swedish
  • Dutch
  • Polish
  • Danish
  • Spanish
  • Norwegian
  • Portuguese

See the full list our full list of supported file types here.

What if I have more questions?

We’re happy to address any further questions or concerns that you may have. Please don’t hesitate to reach out to your Customer Success Manager to continue the conversation.


Was this article helpful?
0 out of 0 found this helpful



Article is closed for comments.

Articles in this section