The algorithms of the protection is based on a locked data set, not just our data set. That's what's in the market right now. It also had good user support. They provide you all the technical details. Additionally, you're able to deep dive into the alerts that are generated. So, you can see which customer has had alerts generated, their other alerts, and it integrates with your existing infrastructure.
Improvements to My Organization:
Once we started adopting these products, we've seen the fraud losses go down considerably. Now, it's not like we also let some customers come in without that protection system, so we don't do a control experiment technically. Since we installed this system, our fraud losses have come down. It could be that hackers are not attacking us anymore because they know that there's a firewall.
Room for Improvement:
From my perspective, I look at models and the methodologies. I think I would like to see them use more of our data, and give us more ability to control or configure how the protection patterns work for a specific business, and also provide more transparency into the methodology.
I wasn't involved in the deployment.
We've had a couple of downstream incidents. However, in general, they're pretty stable. The advantage we get from this product is that, if it's down for a couple of hours, you can store all the log-ins and you can still run it through the system later on. You can back-check.
To be fair, we've adopted them in the last three or four years. The business volume hasn't grown dramatically over that period, but, yes, we've found that it scales.
I wasn't involved in the setup.
I think two main things are important. One is, does it fit in with your existing infrastructure? Second, what are the costs over maintaining it over the life cycle, not just initial costs? Then third is the ability to configure it as your business changes.