Upload your Historical Data

Overview

Learn more about how SEON works and start training our machine learning tools on your bespoke data to hit the ground running when you go live. Uploading your historical transactions to SEON lets you gauge how our tools would strengthen your fraud prevention toolset.

 

Why your historical data matters

Beyond allowing you to test SEON, uploading your historical data lets you get the most out of our solutions from day one.

 

Test SEON's performance

Re-analyze customer interactions and transactions that your existing systems or your team have already seen to uncover metrics that help you gauge how much you could save with SEON. Catch fraudulent activities that slip through the gaps of your current setup, reduce false positive rates, and cut annual review times even during your trial.

Give our Customer Success Team the data they need to shine

Looking through your past transactions will allow our Customer Success and fraud analysts to recommend bespoke rules tailored to your industry and use cases. Why play catch up, when we can give you a headstart in fighting fraud?

Train our Machine Learning Solutions

Unlock the full potential of SEON's machine learning solutions from day one. Improve their already outstanding performance by providing data unique to your company.

 

How to upload your historical data

Use your API integration to send your past transactions to SEON. Beyond being the most efficient and automated option, it also lets you use Label API to train our machine learning tools with granular training data.

If you conducted a batch test before starting your SEON integration, we can upload the results to your profile depending on your preference.

Learning from your past transactions

Past transactions are great for training SEON's machine learning tools before you go live. While our blackbox model will work from day one thanks to the base model, the whitebox model requires personalized training data to achieve peak performance.

Even the blackbox model's performance can improve thanks to your bespoke training data. Given enough data, it will create a unique model for your account to replace the more generally trained base model.

While our machine learning solutions can learn a great deal from transaction states (Approve, Review, Decline), you can get even better results with granular labeling. Create labels on your Settings page. You can then apply these labels to your transactions using the Label API, or manually on the Transactions List or Details pages.

What's next?