What is our primary use case?
UiPath AI Center is primarily used to build an ML model and train it based on invoices provided by vendors. Currently, work is being conducted with a medical firm that has insurance billable invoices, where the goal is to extract information based on business rules. Previously, work was completed for an insurance company where invoices were also processed in UiPath AI Center to extract information.
UiPath AI Center solutions, specifically Document Understanding, are being used in the project, but not Autopilot.
What is most valuable?
UiPath AI Center is found to be very valuable because it has improved significantly in extracting information from each invoice. When an invoice or document is uploaded, it highlights the extractable information and creates columns or labels that can later be changed or removed if necessary.
For each version in UiPath AI Center, a model number has been created. While the updated version is mostly used, the version is updated each time to map it to the model and use it in document understanding for extraction.
UiPath AI Center provides significant time-saving benefits, especially when dealing with documents that are impossible to extract using OCR. Document Understanding offers the opportunity to extract information from challenging documents.
What needs improvement?
Experience with integrating AI models with UiPath AI Center includes a demo for a client where it was implemented in a community edition due to functionality limitations. Required information was successfully extracted using the Document Understanding AI model since the document was quite challenging for OCR to handle.
There is no experience working with the AI model monitoring feature in UiPath AI Center because the community edition does not offer many features.
To improve UiPath AI Center, clients need to be convinced to purchase a separate Document Understanding license because it is quite costly. This is a pain point for clients since they often resort to manual work when OCR fails to extract the necessary information.
Total satisfaction with the functionality of UiPath AI Center is not complete. The platform is improving, but there are still areas that require attention.
For how long have I used the solution?
Overall experience with UiPath AI Center spans approximately one and a half to two years at most.
What do I think about the stability of the solution?
Regarding stability, UiPath AI Center is rated five out of five.
What do I think about the scalability of the solution?
The scalability of UiPath AI Center is full and depends on how the automation structure is designed, granting full power to scale.
How are customer service and support?
Technical support from UiPath was not required as the issues related to classic and modern folder upgrades were not specific to UiPath AI Center. However, responses were found to be general and not tailored to the specifics of the questions.
How would you rate customer service and support?
Which solution did I use previously and why did I switch?
Comparing UiPath AI Center with competitors such as Power Automate, UiPath is superior. Power Automate was found to be less powerful than Document Understanding in UiPath AI Center.
How was the initial setup?
The initial setup process for UiPath AI Center is straightforward and not very complex.
What's my experience with pricing, setup cost, and licensing?
The pricing for UiPath AI Center is generally acceptable. If pricing could be broken down into simpler parts, such as offering a trial on extracting a specific number of invoices, it would help clients understand the ability of Document Understanding.
What other advice do I have?
UiPath AI Center is being used in Document Understanding as a standalone model without collaboration with different platforms.
Integration was not completed through the API. Instead, the extractor was used and the model was referenced as an extraction model to fetch the values.
UiPath AI Center would be recommended to other users. The overall review rating for this product is eight out of ten.
