What is our primary use case?
We use it to automate a lot of data entry or do manipulation in regards to our ERP system.
We are still pretty new with the solution. We have maybe a dozen wizards right now. Some of them are different variants of similar wizards. For the most part, it is used all around our ERP system. E.g., if we have a new product that we're going to make, we take the information from marketing and engineering. That gets handed off to the bot, then the bot enters that information into our ERP. This is a pretty intensive manual entry process.
Everything that we use is on-premise.
How has it helped my organization?
Our company has seen crazy growth over the last 10 to 12 years. We have been up 20 percent year-over-year for that amount of time. Without some of this automation, we would not have the time to invest in developing a full solution or custom applications to handle some stuff because things are changing so quick. It's not worth the investment or there is not enough time to keep up. Constantly hiring more people isn't always an option either. It's sort of a necessity with the growth that we've seen to use this to help with some of the pain points that we have just around growth challenges. Especially at the higher levels of the organization, they see quite a bit of value in the solution.
It has really helped us with efficiency and accuracy. We literally have teams of people just sitting in front of a computers typing in data from Excel sheets all day. That is all they do. We have so many other roles and new challenges associated with our business growth that we could use those bodies. People who have been with the organization a while that we can take those people and say, "Instead of just manually keying all this stuff, go work on something that's a value-add. You know the business and you're familiar with our systems, so go do this instead." That is the biggest win, especially with our state's unemployment being pretty low and filling jobs can be a challenge when finding good people. We want to take the people that we have and leverage them for the things which really help the organization.
It is still a small footprint. In the last 30 days, we've automated 216 hours of work. That is not a lot. However, when you talk about how many people hours that is, it's probably double, if not more than that. So, it is not a huge impact, but in the couple of places that we've deployed it, it has been nice to be able to just send an email in and have it turned around to you from the bot. Within 10 to 15 minutes, the task is done instead of having to send it to a person and rely on them to do it, then get back to you, or they might be on a lunch break, etc. Some of these are critical areas where there's not a lot of time. It's definitely been a benefit to those departments.
I don't think it saves any time from the analyst or IT side. It is an investment upfront to build this stuff and to dig in and find where the opportunities are. We just installed the Kryon Process Discovery service, but we haven't really used it a whole lot yet. Therefore, I can't speak to its efficacy. Hopefully, it will help in this area.
They make it pretty easy to use this solution for the full-cycle of automation from the discovery of processes to turning on the automation and scaling it up. We just installed Process Discovery. In theory, that should work quite well. It recommends what processes if identified are best for automation versus those that are highly complicated or have a lot of bearings. We haven't gotten the value out of that yet because it's brand new.
What is most valuable?
The bot is typically about twice as fast as a person and more accurate, as long as the data going in is good. That is the biggest advantage.
I do like the feature if you find an opportunity that it is very simple to export the recording of that process and import it into Kryon Studio as a starting point to build that automation. It's a really good tool for our business process managers because they can take all that information to their constituents in their departments. Even if it's not robot automation, they have an opportunity to improve the process as a whole. I find that that is an added benefit to this piece of the tool. Just using it as a place to start the conversation and really look at what people are doing on a day-to-day basis. Because managers may be aware of their people's tasks, but might not know the level of manual effort that it takes to do certain tasks. We are finding this helpful. Also, building it in Studio and deploying it into production is a seamless process. It is all very simple and easy to use.
It is just great to be able to traverse software platforms. With all the variable handling and logic that it has, It is an easy tool for a non-technical person to work with and automate across apps. Some of our business people are like, "It's like a giant macro. It works with everything." I'm like, "Yeah, at the basic core, it's a giant macro."
We primarily use it for unattended bots, so it's pure. We don't use any attended bots. So, it's all completely automated using email triggers.
The bot will put in whatever data it receives and it has been really quite effective.
The tool helps with speed to market on anything that we are doing. It is such a quick tool to use if you're doing any sort of automation, even if it's just SQL based.
What needs improvement?
Process Discovery is brand new. We are on 19.1 for it. I know on 19.2 that they change the architecture completely. From talking to some other companies that use this, it sounds like we are missing some pretty big features that we will need. We will be doing an upgrade here in the near term. In general, Kryon's upgrade process is basically uninstall/reinstall at this point. They don't have an easy way to upgrade the software in place, which would be an added benefit. The process is not difficult. We have just a handful of robot machines with high availability enabled, along with a couple of app servers and a couple of database servers. Still, that's 12 machines which all need to be upgraded, and that is no simple effort when you're talking about a full reinstalled software. If we fully utilized our licensing by scaling out, we'd have more than 30 machines. We would have to upgrade on the field. Then, it's a pretty big task if they release new features and things that we want to take advantage of going forward. I see an opportunity for improvement from them here.
We are using it to do some automated reporting, and right now, we can't put images into the HTML formatted body of an email. We can either attach an image or embed a link to an image, but we can't just drop in an image into the HTML. That is feedback that I have given them. It would be nice to be able to have a bot take an image and paste it in as you would in Outlook. Otherwise, we have to provide either those images on a public website. If we want them just to be attached, it just doesn't look as clean. Especially people in our sales force, who are on their mobiles a lot, are not on our network a lot. These are a challenge for them when they just want to be able to glance at the report and go on with their day. It seems like a small problem, but it's limiting for us in some of the areas where we could deploy more of this solution. We have a feature request in for this, and I'm hoping it will be included in the future.
How it delineates file names on email attachments could be better. It is a hard coded comma, and if somebody includes a comment in a file name that messes it up. It is such a ridiculous thing. Who puts commas in file names? But, you would be surprised, and it happens. This is another simple thing they could easily tweak.
Their Tier 1 support is pretty basic. You either have to jump through the same hoops every time or escalate to a different Tier through your rep.
For how long have I used the solution?
We purchased the solution in January.
What do I think about the stability of the solution?
The stability is pretty good. The only unstable part is Studio where you build this stuff. In our version, the 5.25. It has a memory leak. If you leave it open for too long, eventually it just freezes. Saving your work, closing, and opening it just once a day is all it takes. However, this is a bug that I have noticed about the stability of the actual tool. Our production console and robots have been rock solid.
We are still a fledgling operation with it. We have one guy who works part-time on the PDF and another full-time contractor working with the business to build more bots and wizards and also finding more opportunities. Then, there are two of us who are dabbling in it in addition to our regular jobs.
It is also helpful for certain one-off updates. I am a support analyst myself for a lot of our software. People in the business will come to me with really simple problems. E.g., HR had to send out an email to every intern with some of their employment credentials and links to sites for things that they needed. However, they all had to be personalized. Each email needed to have unique information to each candidate in it. So, they built the template real quick. They handed me the email and we dumped a list of the interns from our HRS system on it. We just threw that in the bot real quick with a couple of loops using the email template with some variables and were able to send out an email to all the interns in minutes, which could have taken a whole day, if done manually. For one off things like this, it's such a simple tool and really easy to use.
While there are other ways of doing these type of tasks, e.g., a lot of HR systems have some sort of bulk email feature built-in to send new hires information. Our HRS doesn't have that or maybe it's too clunky to use. This just makes it super quick.
We also use a lot of similar things with one off SQL automation where we could build a complicated script and run that. However, the bot can run a loop of queries for whatever variable it might be and spit out a report. This takes literally minutes to build and run versus hours if you go different paths. This makes it a very sleek little one off thing.
What do I think about the scalability of the solution?
Scaling is simple and very quick. The longest piece of adding any more robots is creating the VM piece and installing all the software that we need onto that virtual machine. Once infrastructure does that piece and hands it off to us, it takes minutes to configure a bot, point it at the server, name it, and get it credentials, then adding it to a production group within the console is very quick. That piece is quite easy.
For part-time use, there are about five of us who are developing it.
There are 50 to 60 people now who interact with the bot by feeding the bot information to trigger it. Also, in our finance space, we have the bot doing some automated work. It doesn't post anything, but it prepares what is called a journal in our system. It prepares a journal, then someone from our finance team will go and review it, making sure everything is accurate before posting that. They are a customer of the bot at that point. That is just a time saver, so they don't have to manually build anything. They can look it over, then they're good.
There are a lot of people who are receiving reports from the bot. I don't know how many there are. It is probably in the hundreds of thousands.
We would love to increase our usage here in IT. We see such value with it. The biggest challenge that we have isn't even with the tool, it is internally funding the tool. The strategy around it is to automate time, then take that time and have a chunk of that time go back to funding more resources to do more automation. It's a self-feeding model and self-funding model. We're still so small that we haven't hit critical mass where we can start really adding more people to it.
How are customer service and technical support?
We have used the technical support a few times. It is pretty standard. Their Tier 1 support asks all the same questions that you always get when you send in a help ticket. Our sales rep is extremely responsive and very helpful with escalating things if we need them to. Once we get past the Level 1 tech support team, then we get some answers.
We are in the process of talking to the technical support a little about doing the upgrade. We personally haven't gone through it yet. The upgrade process is just what I've heard from them so far on how we go about doing it. Once we go through the process and understand it in and out, that will be a big piece of feedback that we will give to our reps.
Which solution did I use previously and why did I switch?
This is the first RPA solution that my company has used.
Someone in the business found the solution. We weren't in a position where we were actually looking for it. One of our business partners stumbled across it, then reached out and said, "Hey guys, this tool looks really cool. We should get a demo going." So, they set one up. Though I didn't participate at that time, one of my colleagues went and participated in that. Then, they came back with an almost like this tool sounds too good to be true.
The tool says it can traverse all apps and can take data from Outlook, Excel, and whatever ERP you have in the web, then put data in, take data out, and manipulate it, or whatever you have to do. They say it can do everything. It was like, "Can you imagine if that's accurate? What we could do with this thing? There's so much potential."
We have grown so fast and our processes have not evolved with the company. We haven't caught up there yet. So, there was a lot of manual stuff going on that we can use the tool for, and we were like, "Wow, okay. So, we definitely have to keep looking at this thing, "and that's when we got our hands on the demo.
We might have been the first company who used the demo once they published it on their website. I remember talking to them after we purchased it, and they're like, "This demo is brand new. You're the first people that we know of to to actually use it or call us and ask questions about it." From there, we did a little proof of concept with it to make sure it would interact with our ERP, then we were off to the races.
How was the initial setup?
The install of RPA was very straightforward. The Process Discovery service install went well, but configuring it properly took a couple of iterations of calls with Kryon to ensure that we had it done right.
We had QC stood up in half a day of production, not including the virtual machines. That wasn't on our team, but the actual install and configuration of Kryon was about half a day. We had two or three calls scheduled to install and configure everything else for RPA and production.
We wanted to make sure that we had a development and QC environment, then production environment. The only strategy that we had around it was to make sure that we followed the best SDLC practices that we could, as far as testing, not coding and prod. One of our guys just set up a call, then we went and did it.
The bot took a month or so to stand up and test, then vet with the business off and on versus the six months that it would have taken to actually build a full solution. Then, if something had changed with any of our systems, that time would have been lost. That is a big savings too, just the opportunity to use this instead of dedicating a whole team of resources to a project like that.
What about the implementation team?
One of my coworkers did the initial setup. We worked directly with Kryon to install it. The resource that they assigned to us was extremely helpful. Even though they are out of Israel, they were very accommodating with our schedules and always willing to help answer questions. They were very helpful and good.
We just needed one team member to deploy it and Kryon did it with us, just one on one. That was a really nice thing. I don't know about the RPA space, but a lot of times when you work with some of these bigger companies, they don't want to support their software or help install the software. They hand you off to a partner. It was really nice just going right to the guys who work to build this stuff and support it. That was a great advantage.
What was our ROI?
We are almost to a point where if I scale out what we're using it for today to the end of the year, then we have paid off the software.
We are probably saving $6000 to $8000 a month in labor costs. It's not a ton, but we're right around $60,000 to $80,000 a year with our current deployment.
We haven't looked into it yet, but for our next opportunity, one of our business partners came across what he thinks is some significant savings in one of our departments. He is really excited to get in there and start looking at the opportunity and evaluating it for viability. If it's good, that'll be a huge win for us. They're talking about six to eight people that we can displace with the bot and and reallocating to other things.
What's my experience with pricing, setup cost, and licensing?
They license per robot and we have 25 robots. You get so many robot licenses and can allocate those licenses to either attended, unintended, or Process Discovery bots. If you have 25 licenses, for example, you could have 20 automation bots and five process discovery. This allocation is flexible. There are also maintenance costs.
Which other solutions did I evaluate?
We also looked at Blue Prism. We just briefly looked at some demos, then we even got pricing, but we were sold from the get-go with the Kryon tool because it was so inexpensive and easy to purchase. It was very easy to learn. We were self-taught before Kryon actually came onsite and did any training for us. We had the demo for a month or two. We played around with it. The demo was very easy to learn, which was a huge benefit for us because then we didn't need to hire an expert who knew the tool already. We could just go and use it.
Kryon is a fairly new company comparatively versus the other players in the space. There are still just the learning too.
Kryon is the most simplistic solution and the easiest to use, but the con that I've heard from other people is that the competitors are, in general, more powerful tools. However, with everything, we've thrown at it so far, we haven't found a use case where it hasn't been able to accommodate, which has been pretty impressive to me. I thought maybe we'd hit a brick wall because it just doesn't have that function or logic to do something, but it's pretty amazing what simple, basic bot programming can do with some logic.
What other advice do I have?
I would really recommend the tool. I definitely see a lot of value with it. Just in the few release notes that I have read for some of the versions they've released since we installed it, it sounds like they are still adding a lot of features and functionality. We are very excited to get this upgrade under our belts and see what new things the new version brings to the table. It has been a great value-add for us personally and easy to use. We are happy with what we bought. No buyer's remorse here.
The biggest challenge that we have had isn't even with the tool. It is with the business funding it. Make sure you have a good funding model set up for funding. Our model is fine if we could actually get more funding for it. That would be the biggest thing. If we could get the business to get behind it a bit more, giving us one or two headcount to really go after stuff.
While it is a pretty simple tool to use, keep all the management centralized in IT, e.g., the operations side of the robots and keeping production access limited to a Kryon administrator who sits in the IT function. That is something that we have done really well. It just helps the stability of the bot. You don't have to worry about anybody going in there and poking at production code or breaking a process. If we did it again, that would be something we would definitely do. Once again, the challenge with that is you need to be able to fund enough people to do that.
We haven't had a lot of challenges with the tool.