Vendor Upate: SEON

March/April Product Release Notes

Vendor Update: SEON

March/April Product Release Notes

Overview

We're rolling out a collection of new Default rules that will help you fight four typical fraud types without in-depth customization. The 97 new rules are divided into four rule categories and designed to catch fraudsters targeting your organization at different stages of the customer journey.

Use case-specific rules

Fraud can rear its ugly head whenever someone interacts with your site. We've designed these rules with this in mind to counter its most common occurrences.

  • Registration rules: It's best to stem the flow of fraud at its source. Block fraudsters and bots from registering fake accounts for multi-accounting, abusing bonuses, or placing fraudulent orders.

  • Login rules: Account takeover attacks are a major risk for you and your users – and both of you lose out if fraudsters are successful. Spot suspicious login attempts to stop account takeovers, avoid chargebacks, or the fraudulent transfer of funds from user accounts.

  • Payment rules: Avoid fraudulent payments, friendly and family fraud with SEON's purpose-built default rules. Blocking suspicious payments can reduce chargeback rates and helps you avoid the cost of handling fraudulent orders in general.

  • Merchant rules: Designed for our partners in ecommerce, couple these rules with those in the Payment rules category to improve the protection of your online store. Catch suspicious orders or merchants before they damage your business.

Getting hold of our new default rules

If you're new to SEON, you should see the new rule categories on the Default Rules tab of the Scoring Engine. If you've been with us for a while, reach out to our Customer Success team and they'll be more than happy to add these rules to your account.

To avoid any issues and conflicts with your existing rulesets, these rules are turned off by default. Flip the toggle switch on the left side of a rule's name to turn it on in your account.

Overview

Fraud prevention can be complicated, but it doesn't have to be. That's why we're excited to announce our integration with OpenAI's GPT, making machine learning rules even easier to understand.

Making complex rules simple to understand

We get it. This isn't immediately easy to interpret:

cpu_speed <= '2283.699951171875' AND only_domain_profile != 'true' AND phone_api_google_exists = 'false' AND region_timezone IN ('+02:00','+03:00','-08:00','-07:00') AND user_label_residential_status IN ('LivingWithParents','Homeowner')

By applying GPT, you no longer glaze over:

User has low CPU speed, is registered on external sites, has no Google Phone number and is from timezones with residential status LivingWithParents/Homeowner - suspicious activity.

When the Whitebox algorithm generates a rule, we use GPT to make it make sense. From Boolean Logic to simple English, you'll know exactly what each rule does and how it arrived at a risk score.

Your data is safe with SEON

And don't worry, your data is safe with SEON. We redact all personally identifiable business information before sharing it with GPT (operated by OpenAI), protecting your privacy.

How to get started

Simply login and head to the Machine Learning tab of the Scoring Engine. You'll see how your existing machine-learning rules are now even easier to understand.

Reply

or to participate.