Automation Anywhere (AA) Review

We are getting some incredible straight through processing rates with AR using cognitive computing


VIDEO

What is our primary use case?

Our primary use case is anything that is structured data and highly repetitive. We also see a lot of gain right now in paperwork, so AR, AP, and record to report. The cognitive product seems to be doing a really good job, and the pace of that going fantastic.

We have done a little bit of everything: 

  • Oracle
  • Salesforce.com
  • Using Citrix, we did an order to cash.
  • In the IQ Bot world, we seem to be doing a lot of AR and invoice-to-pay (those are the two that we see the most). 

We will be moving into medical billing next. Since we have done some really good prototypes with medical records, I think we are ready to go there next. 

How has it helped my organization?

One of the largest use cases that we have about a $10 billion company which does something like 20 thousand AR invoices a month. They had a solution in place. We actual replaced that IQ Bot, and we are hosting it out of San Jose right now. We are getting some phenomenal straight through processing. So, we took what would have taken 30 to 35 people and are able to do that little bit more accurately with about six.

Believe it or not, the first client that we ever did was a Citrix client. We cut our teeth on Citrix around the RPA product. It has some quirks to it, but it is becoming much better. So, we have quite a few instances of that Citrix machine in production at the client. 

What is most valuable?

The most valuable feature of our solution is the cognitive computing because that is coming onto stream a lot faster than you would think. What used to be in the middle of, "Can this be done or not?", is becoming, "It can be done." We are getting some incredible straight through processing rates with AR around healthcare and financial services. 

Integrating RPA and cognitive with other solutions is relatively straightforward. Originally, it came out that Automation Anywhere did screen scraping, but the technology is way beyond that now, where you can do object cloning, even moving into API. We have even coded a couple of APIs into the applications that most people will use. We actually do that now instead of going through a screen. 

Our impressions of IQ Bot are that it goes relatively fast at the pace that it is being advanced every six months, which is good. I believe that we have a couple of the largest installations of it in the U.S. right now. So, we are processing somewhere in the neighborhood of tens of thousands of AP and AR invoices. We actually have robotic operation centers on the West Coast and the East Coast. 

What needs improvement?

The IQ Bot has room for improvement. It is not that it doesn't do a lot of things today, it is just that it can do so much more. I think that that product is going to go crazy.

The next big thing that you are going to see is the analytics engine. I think people finally realize that you can do analytics on what the bots are doing. I think that you will see that come alive in the next six to nine months.

What do I think about the scalability of the solution?

If you write them correctly and have good architectures internally, then scalability is relatively straightforward. This also applies to reliability. However, you sort of need to write a few before you get to this point. As more of these go to production, these are going to be the two big drivers.

Probably one of the faster scales that we have done is we did a prototype with IQ Bot that had about 500 invoices and a 1000 vendors. In about two and a half to three months time, start to finish, we went from the original amount to 15,000 invoices and 2500 vendors. I would suggest in that 90 to 120 day time frame, if you have done things properly and your architecture is good, that you should be able to scale at any pace that you care to.

How was the initial setup?

We have seen both straightforward and complex initial setups. If you are behind a firewall, it is relatively easy to setup. When you are hosting the application, that is when it is a little more complicated (we actually host applications for people). You need to talk to the IT department, as you are actually outside of their firewall, then you are coming back inside their cloud. 

Depending on the architecture, it can be relatively straightforward, and it can be put up in just a day or two. Whereas, if you are hosting it and coming back in, it is a little more complicated.

What was our ROI?

What we did was we took the Automation Anywhere ROI calculations and added quite a few things to it. We actually ask 15 to 18 questions around process time, number of systems used, how people, and how many hours. Ultimately, we can stack rank those ROIs on a process, then we typically try to start with the largest ROIs first. So far, with what we have done, we have seen has been relatively close (as it is almost impossible to hit the ROI exactly), but the ROI on the product is good.   

What's my experience with pricing, setup cost, and licensing?

No one is balking at the cost. They are market rates and will change, but no one is saying, "That is ridiculous or impossible." They are saying, "Show me the ROI and prove that the cost is accurate."

One of the things that will be good for the community will be to get the free version to download, which came out just a month ago called Community Edition. So, you can actually download it and try it for yourself. 

What other advice do I have?

Ease of use is something that you need to get trained on. Once you have gone through the proper training, it is relatively straightforward to use. There is quite a bit of online training on their website, as well as in the Community Edition. After 30 to 40 hours with it, you should be pretty good with it.

One of the promises that we made to Automation Anywhere when we started was that everyone in our company would be trained and certified. Even as a Board Member, I am certified. 

Bot creation is relatively straightforward. Probably within 20 hours, you should be pretty good at it. There are some nuances around it though that I think separate first time developers versus seasoned developers. What you will see in the marketplace is how easy it is to develop one compared to how easy it is to develop one that runs in production 99.9 percent of the time.

Across the board, I would give it a nine out of ten. There are some areas that they can move up into, but they are working on it right now. I am assuming that everybody is working on it, so we will stay in that nine range.

Go do your homework. Call some of the larger clients because they seem to face the largest problems. Automation Anywhere has five to ten gigantic accounts that you can just talk to. Ultimately, I would just be careful in the way that you implement it. It is like any tool, if you know what you are doing, "Great," and if not, then you can make a mess. Therefore, be careful and know what you are doing.

Disclosure: IT Central Station contacted the reviewer to collect the review and to validate authenticity. The reviewer was referred by the vendor, but the review is not subject to editing or approval by the vendor. The reviewer's company has a business relationship with this vendor other than being a customer: Partner.
Add a Comment
Guest
Sign Up with Email