I have used UiPath Orchestrator, and we have created both attended and unattended robots for our clients.
We have been using the new AI and OCR technologies with UiPath, and we are currently trying to implement the Citrix log capability that was recently introduced.
We are not running our automations in a virtual environment. When we automate any Citrix-based application, it's all email-based. There is a Citrix receiver and we communicate with that, which helps automate Citrix applications much faster.
Most of the clients I had seen have been running in virtual environments, although I have seen some of our clients running on the desktop. We have also seen hybrid scenarios.
One thing is that virtual environments can be standardized pretty quickly. So, that's an advantage. Normally, the companies, which are leaning towards more cloud now, will be happy with this. So, I think that is one factor. As you move virtual machines to the cloud you can migrate your bots to the cloud faster.
I have worked on various different domains including the public sector, commercial, healthcare, energy, utility, and federal. These are the different customers for which we are implementing solutions. Now, the customers are moving towards AI and natural language processing. They are more into chatbots, how they can use artificial intelligence, making use of data science, and putting more machine learning on board.
With respect to how easy it is to automate our company's processes, on a scale of one to five, I would rate it two and a half. I'd say it is about marketing. You can develop anything. There are very small processes that you can develop with having minimal experience. However, when you start implementing complex processes, I would say you need to be a background developer.
On a scale of one to five, judging how beneficial it is, I would rate the training a five. All of my team members have been using UiPath Academy for training and certification. It's not just with the U.S., but outside the U.S. as well.
From the point that a UiPath license is purchased until the first robot is ready totally depends upon what use case we are implementing. There are different methodologies that people use. Some build the bot without exceptions and it can go to production. Like a very simple process can go to production in two to three weeks. A more complex bot will take eight to ten weeks, and depending upon the process, it can go longer. I have seen tasks when a human is performing the job and it takes him around twenty minutes per transaction. But, when the bot comes in, it actually completed that same transaction in five minutes. But, to develop that five minutes of processing, it was understanding system availability and testing. Then you have to do load testing. It takes ten weeks or so.
Our clients decide to implement RPA for several reasons. The first reason, of course, is to have work completed faster. Second, when there is a workload, you can work on it more efficiently and with fewer people. Consider an open enrollment in October, where the open enrollment starts at 10:00 AM and there are a lot of transactions flowing in. Now you have to hire a human and train them. With the bot, we can just scale up instead. Finally, the bots are errorless.