What are some areas in Acumatica where predictive analytics or true AI could help with managing administration, helping users with repetitive tasks, or generally making the system more valuable? Some of this could be done with notifications and business events, but I’m think of something even more advanced. (Ex: Month-end close? Notifying users of situations like low inventory? missing tasks? Unfinished activities? Orphaned transactions? Bank rec? Import or integration management? Case management? Customer/Billing management/)
Where Could Predictive Analytics Help in Acumatica?
Userlevel 5
+1
I believe in 2022 R2 they implemented a GL anomoly detection tool which creates a probability score for each transaction as an anomaly or unusual activity in the Acumatica GL environment. I believe there is a pre-trained model as well as the option to train your own model using a company’s historical data.
I think they also have some machine learning models deployed to identify system issues.
The internal moving average forecast model is probably “good enough” for most businesses regarding low inventory. We identify all items with less than a month’s supply on hand using a 12-month moving-average model. If you need something more serious than that, there are external forecasting tools.
I personally feel that they should focus on the GL anomaly detection tool for now and flesh it out to do classifications (defined by the user). I believe more than one end point would be valuable (i.e. overpaid, underpaid, duplicate transaction, wrong journal destination, etc.)
Userlevel 4
Thanks
Reply
Enter your E-mail address. We'll send you an e-mail with instructions to reset your password.