- Write manual rules in plain English within minutes, test their efficacy on your own data, and implement them immediately into your fraud model
- The manual rule builder provides signals to the machine learning engine, which in turn detects pattern to be implemented back in the manual rule builder, so every facet of Simility’s system augments the others to fight fraud
- Simility’s state of the art machine learning engine adapts, strengthens and quickly reacts to detect new fraud schemes
- Easily source, transform and consume unstructured data in any format–such as emails, images, and transactions–and combine it with Simility’s vast repository of known fraudsters
POWER UP YOUR MANUAL RULE BUILDER
What is fraud analytics? Most fraud models consist of hundreds of manual rules and logic statements that describe behaviors associated with fraud. Traditionally, analysts had to write complex statements or SQL queries to encode these rules. With Simility’s manual rule builder, multiple data transforms are precomputed so anyone can quickly and easily write a simple rule that captures extremely complex logic and test it on your company’s own data. The fraud-fighting power of your rules engine will be in the hands of the people with the most experience and intuition in detecting fraud–your own analysts.
MACHINE LEARNING AT ITS BEST
Simility allows analysts to create signals and manual rules that codify their insights, which are then fed into Simility’s machine learning models. Through big data predictive analytics, your fraud model will evolve based on every decision and rule, continuously enhancing the detection of anomalous patterns. This means your model improves its defenses while you sleep. Simility’s machine learning engine begins with experience from an array of business verticals, then quickly adapts to your unique needs. It detects patterns of fraud before they are perceptible to human analysis or become a problem, keeping you ahead of the fraudsters.
Any piece of data can be useful in detecting fraud and abuse, but not all fraud detection solutions let you import all types of data into your fraud model. Simility accepts unstructured data, so you can programmatically detect patterns in ambiguous pieces of information, such as the tone of an email or the appearance of an image. Your predictive model is built on top of Simility’s foundation of known fraudsters, so you get the advantage of our extensive experience from day one.
INSTALL FREE Install in seconds. No credit card required
- Identify unique devices–desktop or mobile–by their fingerprints
- Use device risk scores–computed using proprietary algorithms–to determine fraud probability for any device, even one that has never been blacklisted before
- Determine whether two different device fingerprints–highlighted by clustering algorithms–are likely to belong to the same fraudster
- Get complete access to the device attributes for use in your custom fraud models
HOW IT WORKS
Device Recon analyzes hundreds of mobile and desktop device characteristics and behaviors–including browsers, language, location, operating system, even mobile emulation, and battery level–to fingerprint devices. Fraudsters can mask identifying properties like their username, email, and IP address, but with Device Recon you can determine if a device is associated with fraud. If a fraudster returns to your site or application and makes minor changes to device characteristics and behaviors, traditional fingerprinting technology will not detect it’s the same device, and the fraudster will slip through your blacklist defense. But Device Recon uses machine learning models that incorporate risk scoring and clustering to see through the fraudsters’ techniques–giving you triple protection against their attacks.
Our powerful machine-learning models determine the riskiness of every device, even if it has never appeared on the network before or is not on a blacklist.
To calculate the probability that a particular device is being used by a fraudster, we deploy our proprietary algorithms against an ever-growing knowledge base of external data points, such as IP belonging to a Tor network or likelihood of a browser version being used in a particular country.
A device fingerprint captures the state of a device at a given point in time. But device characteristics–even those of legitimate users–naturally change over time, such as when a plugin is installed or operating system is upgraded.
So Device Recon deploys a second machine learning model that uses statistical clustering to distinguish between major and minor device changes, which enables it to determine whether two device fingerprints actually belong to the same device.
- Save time with our unified interface. Manual rules, machine learning, and Device Recon data are displayed in a single analytics view so your analysts don’t have to toggle between data analytics interfaces to see the whole picture. Simility simplifies the fraud screening process.
- Stay on task. Create and deploy custom workflows in minutes. Analysts stay efficient by automating their own work.
- See things clearly. Use sophisticated data visualization with slicing and dicing features to identify patterns and relationships and screen for fraud.
Simility pulls huge volumes of data from disparate sources, and the most relevant signal is presented to analysts in a single interface so they see the signal without noise. Manual rules, machine learning and Device Recon data are all incorporated into a single fraud profile for each user. The signals from one data source inform the others, so analysts no longer have to click between views to make decisions.
Analysts can customize every aspect of their workflow to show the most relevant data, dive deeper when needed, and make bulk decisions to optimize efficiency. They can even automate their own workflows. The intuitive fraud screening interface lets them test and revise fraud scores and risk thresholds to match the specific profiles of your customers, ecosystem and business. Automated workflows facilitate faster analysis and better decisions that defend your data and your business from fraud.
Our graphical network analyzer is a rich display that shows all the connections between various data types and allows you to expand and explore the complete network. Analysts can quickly slice and dice data along any parameter and accounts of interest can be isolated in bulk. Our advanced reporting capabilities let managers track the health of their entire system at a glance.