Settings page

The Settings page allows you to adjust and finetune how your SEON implementation works. Customize the interface and rule settings to meet your needs and fit into your workflows.

The page is divided into four tabs. Each section contains options connected to various elements of the SEON system. 

• The General page houses timezone and locale settings, along with default expiration timeframes.
• The System tab provides a host of options to finetune how rules work.
• Authentication houses all settings connected to Single Sign-on (SSO) and two-factor authentication.
• Machine Learning is where you can customize how AI can help your work

On the top, you'll find basic account settings, such as the time zone, locale, base currency, and default IP expiration. 

Timeout settings

API timeouts define how long you can wait for a response from SEON's data enrichment APIs. A higher timeout improves accuracy but leads to a longer response time. We recommend only using these functions at the earliest user access point, e.g., user registration. 

The minimum timeout is 1500 ms, but we recommend using a timeout of at least 1700 ms.

Learn more: Head over to our guides on using the Priority timeout and desired accuracy settings.

Flagging settings

On the bottom, you can enable/disable flagging and auto-flagging blacklisted values by using the toggles. 

Changing thresholds

Fraud scores are based on default, custom, and machine learning rules. You can review the different rules in the Scoring Engine. By default, a score over ten is considered risky. However, every business has a different risk appetite, so you can adjust the different state thresholds to your needs. The default thresholds for each state are:

Use the slider to change your threshold settings:

Note: We tailored these thresholds to high-risk merchants. We recommend you adjust the DECLINE state threshold between 20–35, depending on your experience with SEON in the first few days, especially if you encounter any unwanted false positives.

Adjusting weight

In SEON, Fraud scores are the sum of the Email, IP, Phone and custom rule scores. If you need to emphasize the importance of one category over another, you can adjust the weights applied to the Email and IP scores.

For example, if you set the IP weight slider to 75%, the fraud score applied by each IP rule will only be three-quarters of the default setting.

Note: Keep in mind that you can also use Custom Rules to change the Fraud score.

State conflicts

Some rules in SEON can change the state of a given transaction without directly changing its fraud score. A transaction may trigger multiple state rules with different effects in some cases. Set the preferred outcome of these conflicts using the selector.

Single Sign-On

SSO login can improve the security of your SEON accounts by requiring your team to sign in through an identity provider of your choice, such as G-Suite or Azure AD. Learn more about setting up SSO, and contact our Customer Success Team if you run into any problems.

Two-factor authentication

Two-factor authentication – or 2FA – improves the security of all SEON accounts by adding an additional security layer when signing in. 

Members of your team will be required to provide a single-use passcode generated by an authenticator app on their phone when signing in. 2FA can be used to counter account takeover attempts by requiring a second device to access SEON.

Labels

Set up positive or negative labels to train SEON's Machine Learning system. Labels can be set automatically and sent to the system through the Label API or applied to transactions manually on the admin panel by clicking on its state (APPROVE / REVIEW / DECLINE). The system is tied to your account and only trained on your data.

Use labels to define the reasons why you accepted or declined a transaction. Negative labels can mark why a transaction was considered fraudulent (e.g., chargeback, bonus abuse, stolen card, etc.). On the other hand, positive labels highlight proof of a valid customer (e.g., passed KYC).  

You can choose to only train our Machine Learning algorithm with the feedback you provide in the form of labels by clicking this switch.

During manual reviews, you can label transactions "Marked as decline" or "Marked as approved" or any other negative or positive label you have previously set up.

Complex rules

SEON's machine learning system automatically generates complex rules. You can choose to automatically apply these rules to your transactions if their accuracy is above a certain threshold here.

Use the toggle to allow REVIEW or DECLINE state rules to be applied. Adjust the accuracy threshold with the slider.

Note: You can turn any complex or heuristic rule on or off manually in the Machine Leaning Rules tab of the scoring engine. 

Heuristic Rules

When a negative tag is applied to a transaction, SEON will link the data to other previously declined transactions in your account. Heuristic rules block all future transactions from the same data point. 

You can choose to turn on heuristic rules above a set accuracy automatically. Use the text field to specify which data points our system should monitor. Finetune the minimum accuracy using the slider.

Note: You can turn any complex or heuristic rule on or off manually in the Machine Leaning Rules tab of the scoring engine. 

Blackbox Score

The blackbox model is only available for customers with at least 1000 transactions where a minimum of 100 transactions are in the declined state. Learn more about SEON's Blackbox model.

Click to toggles to finetune how you use Blackbox scoring:

Adjust how AML API works for you to make sure your teams has all the information they need to keep you compliant and fight fincrime.

Monitoring Settings

AML API checks names against lists in four categories: Sanctions, Watchlists, Crimelists and PEP lists. However, each category bears different compliance risks. Use these settings to adjust how often SEON's automatic monitoring tools should recheck names that you've enabled automatic monitoring for against each list type.

Why change monitoring frequency? Different jurisdictions require varying levels of vigilance from organizations to ensure AML compliance. However, constant updates, especially in the case of common names, can cause a lot of noise for Compliance Officers in the form of false positives. Adjusting these frequency options helps you stay compliant while minimizing your teams' workloads.

Frequency options:

• On every database change
  Monitoring checks will run every time SEON's AML databases are updated, this can happen several times a day.
• Daily
  Checks will happen once a day.
• Weekly
  We'll run monitoring checks weekly based on when monitoring was enabled for a name.
• Monthly
  Monthly checks will run on the day of the month on which monitoring was enabled for the name (e.g., May, June, July 20)
• Quarterly
  Quarterly checks will happen once every quarter, or every three months from the day monitoring was enabled (February 11, June 11, September 11).
• Twice a year
  Monitoring will check this source type once every six months, using the day your enabled monitoring as the base (e.g. June 15, December 15)
• Yearly
  Yearly checks will run every year on the day you enabled automatic monitoring for a name.