What is our primary use case?
Our primary use case is two back-office processes from a specific area, an area which was very complicated. In the credit card world, there is a process called a chargeback, when the customer says he didn't do a specific transaction, or it's not the right amount. This process has a very long, complicated and regulated process behind the scenes, after the conversation with the customer. We used a robot to automate a lot of the process, to make it shorter, easier, and to save a lot of agents working on the process.
Every week or two weeks we are bringing on a new process with the robot. We are still in the middle of the automation move; we have a very big roadmap for what comes next. We are still on this trail.
How has it helped my organization?
The process I noted above is a main business area with a lot of very specific processes inside - we have now more than 15 specific processes in this business area. All of them, as of now, have saved us almost six FTEs. That's a lot of money. The savings are great and they will increase every month.
In the last month, we achieved 11,000 activities done by robot. Before Kryon, we did 500 each month. When we started with automation ten months ago we did a few hundred. Each month we have added more and more processes and more and more activities. It's quite dramatic, because we took a very complex business area, and every activity in it which can be done automatically is being replaced by the robot instead of people working on them.
In terms of Kryon helping us embrace digital transformation, we are very advanced in this area. More than 70 percent of our interactions with customers are digital. But Kryon helped us with some of the processes which are more complicated, where we had people doing things in the back-office after the digital process. Now we can do more of them automatically. It's another step in the same direction for us.
What needs improvement?
What we are lacking is OCR, the ability to read text. We would like it to have the ability to take a page which is free text and analyze it, and then for the robot to know what to do next. This ability does not exist in Kryon products, so our company is now trying another tool which has this specific feature - an IBM tool which combines RPA and OCR together.
Also, when the robot stops working for any reason - it doesn't matter if it's something in our systems or our infrastructure - it does not know to go back to the same point it was at when it stopped. We have to start from the beginning or to delete the report. It cannot continue from the same point. We would like the robot to know how to continue.
For how long have I used the solution?
We launched in October 2018, about ten months ago.
What do I think about the stability of the solution?
It's quite stable. Most of the difficulties are coming from the changes on our side.
What do I think about the scalability of the solution?
The scalability has been good enough for our needs.
Which solution did I use previously and why did I switch?
Prior to Kryon, we didn't work with another solution. But we are now starting to work with another solution because of OCR.
How was the initial setup?
On the technical side, the setup was done mostly by HMS. The issues we dealt with were mostly to do with the connection to our organization's system. The authorization issue was difficult for the system and many other aspects, but I know part of it was from our side. We also had some information-security issues. I don't think it was too complicated, but it took us about a month-and-a-half, to do the first at set up, and then it was much easier.
We learned a lot. We learned what we need to prepare for each process so the process will be the most effective and quick. We learned about the abilities of the tool and that helped us in thinking about the next processes and how to do things more efficiently.
We are also using BI to help us with more complex business logic. Instead of doing it in the robot, we are doing it in BI and giving the robot more direct reports to work on. It makes our process much smarter and efficient.
In terms of deployment and maintenance, we have one full-time employee working on it, and one of our business persons, a project manager, is working on it half-time. He's involved in building the business process, deciding what will be the next step, and helping close all the details on our side.
The maintenance is not high. It's just that each time the system is changing, we need to teach the robot the process from the beginning. This is the only maintenance we have. The other work is really to in developing the next process each time.
What about the implementation team?
We have a supplier doing the implementations for us, HMS. They are more involved in the technical details and issues. But next month, we are planning to do Kryon's course. We are bringing it here, and then eight people, both business and tech people from our organization, will learn the tool and then we will be able to use it by ourselves.
Our experience with HMS has been very good. We like working with them. We are continuing with them but from a budget point of view we would also want to learn to be more independent. But we'll definitely continue working with them.
What was our ROI?
Now that the process is cheaper for us, we have been able to change our business rules a little bit and save more money by automating. To give a simple example, if it was worthwhile for us to do a specific activity only if it was more than 50 shekels, now, when the robot is doing it and it's cheaper for us, we're doing it from the first shekel. We are saving more money by changing the business rules because the process is costing us less money.
It has also helped to reduce employee errors. This is also very important. And it has helped us in terms of monitoring because we need to do less monitoring of the employees. That's another area of savings, another improvement. In the beginning, we did need to monitor the automated process, but after seeing it and becoming comfortable with it, we were able to reduce our monitoring of employees.
What's my experience with pricing, setup cost, and licensing?
The first robot was very expensive.
For us, there are the additional implementation costs that we paying to HMS.
What other advice do I have?
We have a lot of people coming to us to learn from our process, so we've been providing advice already. First of all, it's important to do a very smart business analysis, from the business side, and not automate exactly the way it was done before. Really think about ways you can improve and make the business logic, the business rules behind the scenes, more effective. This is very important.
It's also important to look at becoming independent with the solution from the beginning. We have only started to plan for that now.
The authorization issues within our company also took a lot of time, so I would suggest getting those dealt with in advance.
Part of the difficulty in embracing the tool is on our side, due to bureaucratic issues that we have in our organization. Every change we make which results in a new version of the system the robot is working on means we need to teach the robot from the beginning to do the process.
For example, when we would have a new version of the system, we would ask our people, "Has anything changed?" They would say, "No. Nothing is new." Only after would we find out that something moved from the right to the left. Just one button. For the employees, it didn't matter. For the robot, of course, it does. And then we needed to teach the robot to do it again. So we have learned to prepare ourselves better before a new version is released. We receive all the screens. We teach the robot to do everything again. We're checking ourselves. We test more. These are things we learned along the way, "on-the-job learning." We wasted a lot of time on this because we did not prepare for it properly.
In terms of the prioritization process, we are working with a business analyst to decide exactly what the next process should be, what has the biggest number of activities per month and that the business case is important. We also look, from an implementation point of view, at whether we can do it. Is it accessible for us and not something very complicated which will not be possible? We take all of this into account and we decide which process to put in next. We already did the main business area and now we're continuing to another business area and doing the same process. One of the things on our roadmap is to optimize part of our monitoring processes.
From a business-analyst point of view, we are using a lot of analytics to make it very smart and efficient. So it's not saving us money in business analytics, it's the opposite. But it's worthwhile for us. The discovery process in our methodology is very important because each case is a little bit different, so we need to find the right rules, the logic.
The first process was the most difficult. We needed to learn how to work with it. But by using it more and more, we have achieved a very easy and quick process. The delivery time, now, is very good for us. We are adding a new process every two weeks. Now it's good, but the first time was more difficult.