What is our primary use case?
I don't remember every single use case. One is if you have many fees. I don't know if you know the Intuit payroll system, we acquired it from a company called Paystation. Paystation was pretty good in that it was generating a few hundred million dollars per year as revenue. And that's what prompted Intuit to buy payroll. It's an old application, from the late nineties, when Java was proven to be one of the top technologies for building enterprise applications. We had the state of the art technology, which was used in the early 2000s which is the JSF-JSP server.
I think Kount has the means to check if this guy can actually add a backup on the addition, backup on modification, and backup on deletion. I don't think deletion is such a serious thing, but it includes backup, deletion, addition, and modification. Especially modification considering these two use cases.
Across the application state, you may have payroll integrated with QuickBooks online accounting system. That's a different type of payroll, the way they interact is different, and there is a stand-alone payroll, and for some reason, users love it a lot. They want stand-alone payroll so that the kind of UI pages are different.
There is another use case: Even the agents can do fraud. So for agents, there was a separate system. Across these three different vertical flavors of Interact payroll, the first one was with a few cases about banker forms. The second was about running a payroll. Even with running payroll, there are PMs who have the kind of experience with payroll. They come up with what use cases are most important.
Running payroll is another one. Even in the back-end, there was something that involved Interact, anything and everything about moving money and modifying accounts. Even employee addition is a very sensitive thing. Employee addition and employee modification. Sometimes people simulate that I'm monitoring this employee little by little and there is a lot of fraudulent activity. There can be anything and everything. I had 12 or 13 different use cases when I worked across the three verticals in the field.
How has it helped my organization?
There were other competing systems which were there. There were other people and other departments using other software, but Kount stood out in payroll because it was already proven and even payments division was using it. Kount was the best at the time that I was working with it.
What is most valuable?
I was working with the UI integration part. I have an understanding of what happens in the background as a black-box. I mark second transactions as good or bad and I posted on the back-end for them to look at this information to do the first thing. It's very critical. There are a lot of complaints and government laws around this so if I can't generate my paycheck at 3:00 PM for you, it's perfectly okay. But I want to get the check by 3:30 and if it doesn't happen, if the guy takes a different route, then we have a problem and there are a lot of complaints. We faced something like this in 2051 or 2016, I believe.
What needs improvement?
In terms of what could be improved, that's a very good question. There was this layer you had to go to where I think we needed to add the frame. Then there is one tone for communication that happens with us and Kount and the response comes back again. So the time that is taken to go to Kount and come back should be in the order of around 100 milliseconds or less. And our context was taking around 200 to 300 milliseconds. We didn't want the extra load of 100 milliseconds to happen, so if the two rounds of stability could be cut to one, that would be very helpful. That was very relevant for the mobile app.
For how long have I used the solution?
I was totally engaged in work with Kount for six months. I would say it took about a week or a week plus to get a grasp of it. Then there was some work experimenting which was done for another two weeks to know whether this works or not. So I'd say a month. And then you actually start coding and doing experimenting, that's a month and you work with the QA and then you educate them on what has been tested and how it has been tested. Luckily they got a very very good, skillful person who was very cooperative. So that took about four months to stabilize, and after roughly two months or so we were in production to make sure that everything is okay and it's actually usable by the analytics team, how much fraud we are preventing, and how much money we have paid for the company. It took about six months.
What do I think about the scalability of the solution?
In terms of scalability, I would say that depends on Kount's back-end capacity and what payroll was assigned. If you have 100,000 users using payroll and all of them come on and add an account, everyone invariably goes through the screening process for whether they are allowed to or not. I think it was smooth. Otherwise, we would've had no updates, right? With performance and different aspects, every time a user gets to call the production server there is latency induced because of very poor upgrades, and bad experience. But we didn't have any kind of bad experience reported at all. We didn't have anything.
How are customer service and technical support?
Initially, I had to interact with people. Whenever we used to have any kind of issues, there was this product manager cum technical person and he was pretty good and knowledgeable. I remember tracking quite a lot with this guy.
We had some trouble and there were some misunderstanding. After a few months we had this genius discover that we were doing something wrong, that the requirements were wrong. We were posting duplicates to the code in the backup. This information of the unique string was not unique for them and the back-end had trouble identifying who was doing it, which is some king of mapping issue. I remember working every day for a week or two and that was good. The tech support worked.
How was the initial setup?
The initial setup was absolutely straightforward. Within a week I was fully working and within one day I could set up the demo. I liked that initial experience of setting things up. If data is bad then you have back-end problems and things won't work. But with Kount it is pretty good. I was able to create a demo and then in one day, from the point where I looked at the first URL, I was able to set it up immediately. And they had a QA URL which you connect with and it performs within a day or so.
What other advice do I have?
If it were 2017 or 2018 I would have definitely recommended it. There were other departments in it. A lot of my colleagues said they were using other systems. One name came up frequently that I don't recollect. I wondered why they were not using Kount. There might be some other reasons, though I don't think there is any technical problem. I enjoyed the work. I didn't have any complaints and there were no performance issues except which were for the most part induced by having external problems.
After doing everything, I needed to have some credit for the work. So I used to get statistical data from the risk management team and the product managers were asking how Kount was. They used to say that so many millions of dollars were saved. They definitely came back and gave this feedback. It helped pay a lot of money for a pickup.
On a scale of one to ten I would rate Kount a nine.