What is our primary use case?
We started off with accounts payable type of processes. That was our proof of concept. We built that within two or three months. It was highly successful, then we got a high level of endorsement from the wider organization. From there, we moved into academic processing where there's a lot of opportunity around the processing of student results, assessing students for admission into the university and handing out offers to students. There are a lot of things in that academic space which are highly manual. Often, these take quite a long time to process, so there's a clear benefit to the student by reducing the cycle time, which has led to great benefits for the organization.
What is most valuable?
In my case in particular, I come from a business background, not an IT background. So, it sort of speaks for itself that I was able to pickup its functionality and how it works. In the space of two years time, I've gone from a junior developer to a senior specialist. So, it really speaks for itself in terms of being user-friendly.
What needs improvement?
There are often a lot of functions within the software that will do the same thing, but it's up to the developer to decide which is the most appropriate. If there were a way of the product sort of indicating or automating which function is most applicable for that field for that particular purpose, that would be great. I think that might be a ways away. But, I think a lot of the improvements that we're looking to find are going to be in the new cloud release. So, we are really looking forward to getting on that.
What do I think about the stability of the solution?
Every piece of software has its downfalls. The great thing about AA, particularly in Melbourne, is we have local support. So, if we do raise a ticket, it's addressed by our local team down there. If they can't handle it, it'll be expedited abroad. So, we're always very comfortable that we're going to get a prompt reply and support, if for whatever reason, we have an issue with the software. We've sort of developed coding standards which has helped us to avoid any downfalls or pitfalls. Throughout our journey, we've learned how the software really works. We're developing it in a more mature way now, which has reduced the number of errors and breakages of the code.
it's just a learning process. As long as you're willing to learn and engage with the software, it can be quite stable.
With Automation Anywhere, there are multiple ways to achieve the same outcome; many different ways to design a bot and execute a bot. But, there's definitely a best practice out there. For those with little experience, they might be inclined to follow the quickest or the easiest way to develop a bot, but that's not necessarily the most sustainable solution. So, we sat down as a team at the university and agreed upon a certain set of standards that we will always follow. We initiated a code review process, so every code that goes into production, or every bot that goes into production will go through multiple rounds of code review. If it doesn't meet our standards, it'll be pulled back out of production, reiterated, and eventually then it'll go live.
What do I think about the scalability of the solution?
We've gone down the Reagan approach, which was referred to in the session earlier today, where we essentially started out with that low hanging fruit. We built confidence within the business and from there, it was quite an easy sell because it was quite obvious to the business that we could deliver and could deliver fairly quickly.
In terms of scalability, we've had to grow the team slightly. It's really important to have the right team. You need someone with a lot of experience around the tool. From there, you can bring in a range of skill sets, including SMEs. We have SMEs on the team now that don't come from an IT background or development background. By having that range of skills, we've been able to sort of up the velocity of releasing bots.
We could always do better though, and we're still learning.
How are customer service and technical support?
The technical support in Melbourne, it's great. We have a strong relationship with Andrew Rowlands, who's our customer relationship manager down in Melbourne. We really enjoy working with him. If we do have any issues with tickets that have been unanswered, he will address them for us. I think one of Automation Anywhere's strong points is really its focus on the customer. Because naturally issues do arise with the software here and there, some of which have been addressed in later releases of the product.
We haven't yet updated to the latest release. The company is well and truly able to support all versions and is always really responsive. We typically will receive a solution within 24 to 48 hours.
How was the initial setup?
The initial setup actually came before my time at the organization. I think it is a little bit of an ongoing struggle, because we don't deal with the infrastructure. We have a separate infrastructure team that sorts all that out. So, sometimes there are communication issues just trying to get our internal IT team onboard to run through the installs and that sort of thing. But, definitely with the new cloud version coming onboard, we'll be looking toward going to that. Hopefully it'll resolve some of that friction that comes about through the install process and complexity.
Obviously, the easiest way to create a bot would be using things like keystrokes. It's very intuitive, simply entering the database and what you would see on a keyboard. However, it doesn't really relate to what you're seeing on the screen. So, quite often, you'll be entering data in the wrong field if there's a performance issue with the system. While if you use object cloning, which is another function within the software, it's more dynamic. It'll wait for certain fields to appear, then you will lower the risk of breakages of code.
What was our ROI?
Just an example, in a period of two to three years, we have a positive ROI. We have a positive return. We've delivered 25 to 30 bots, which equates to over 1000 atmx files. So, for a relatively small team, we've been able to scale up quite effectively.
Which other solutions did I evaluate?
In regards to how AA works, UiPath has more of a design flow view, where you'll be creating process maps and executing bots face on the process flow. While Automation Anywhere has this task view, or line by line view, that I find it very easy to use and intuitive. It's just really a great piece of software.
What other advice do I have?
It's a solid eight out of 10. We've delivered tremendous value in a really short amount of time. We've gotten a lot of support from our organization. Of course, there are always improvements which could make it even more powerful. But I know the company is well and truly across those. I'm sure it's going to go from strength to strength.
From the business side, we're not an organization that has tremendous volumes, like a big global bank, perhaps that sort of comes into play a bit. But again, it comes back to coding standards. We have a lot of confidence in those standards. We just really ensure that new members of the team are strictly following those for sustainable bots.