What is our primary use case?
Our core business is BPO, which is Business Process Outsourcing. We have massive operations that we have to perform for our customers and we have a digital section of the company that is assisting with that. The digital section is relatively new, being no more than two years old. We are building a number of solutions and tools that our digital section is using, and RPA is one of these tools. The goal is to help our customers innovate and assist them with their digital transformation, ultimately making them more efficient and more profitable. This is possible because some of the processes are very repetitive and performing them with humans is a very bad choice.
We have a hybrid environment, where some of our functionality is on-premises and some is on the cloud. For example, we have some cloud-based automation, and we use UiPath Apps, which is on the cloud.
We had a successful use case at the beginning of the year where we needed to process a large number of invoices that had contained errors when they were originally sent to the customers. There were approximately 200,000 invoices and we had a deadline of four days to complete the task.
It began with us developing the bot, which was completed in less than a day. After that, we sent the bot to our production environment to start processing the invoices. We were successful in the task, through the parallelism of 50 robots, we could process 5 invoices per second.
We have some metrics that describe how long it would take the process to be completed manually. It takes a human an average of between 60 and 90 seconds to process just one invoice. We estimated that it would have taken approximately 125 days to complete this task manually, with between 250 and 300 people working on it together.
How has it helped my organization?
We don't use the low-code functionality. Rather, we use the typical development features. When you're talking about developing inside the UiPath, you have something very user-friendly, so you don't even need to use the low-code options. It is very intuitive and you don't need to know technologies such as C# or .NET to develop automations.
The use of UiPath has helped to increase customer satisfaction by a lot. Our main goals are to improve the average handling time that the customer needs to complete transactions, as well as to improve quality. Customer satisfaction improves not only with the financial benefit resulting from a better average handling time but also, from improved quality in transactions. Our human resources department uses tools such as surveys to investigate the quality and they have their own metrics and KPIs for customer satisfaction.
Our first-contact resolution rates have increased because as we develop successful cases and implementations for different customers, future customers benefit from this through faster service, which leads to better customer satisfaction. I estimate that our first-contact resolution has increased by 15% to 20%.
The time it takes us to create automations depends on what we assess the complexity of the bot to be. We have a methodology and metrics that have been developed by our Center of Excellence, categorizing the bots into small, medium and complex. A small bot, which has simple logic, is something that we implement in between one and two weeks. A medium-complexity bot has a timeline of between two and four weeks, and a very complex bot takes four to six weeks to implement.
What is most valuable?
The most valuable feature that we are using is UiPath Apps because it makes it very easy to implement tasks. It is very easy to scale operations, which is important because we're not talking about just five or ten agents. We're talking about 1,000 to 2,000 agents. The Apps feature helps us to scale very quickly and very easily. We only need to develop one or two bots and then link them to UiPath Apps to process everything. All of the integration between the bots and the human, along with any scheduling that needs to be done, is taken care of by Apps. In our situation, the Apps feature is the best solution to handle this scale.
Utilizing our bots is very easy, and it is done using the licenses that we have with partner UiPath. We can access our licenses, then distribute them to the customers and we can use them dynamically. This is all done in a very easy manner. We just have to navigate to the web-based hub, where we have access to everything that we need.
UiPath is highly customizable and this is helpful for us because we can develop models and frameworks that can be reused for different tasks and different customers. For example, if we have a customer with a process that is very similar to one that we have previously developed for somebody else, we can reuse the models to scale the bots. This makes the new development very easy and very fast.
The Agent Console is able to provide customer insight in conjunction with the task and process mining features that we use. We install the tool into the machine that the customer uses every day, where it will capture the manual tasks and processes into a database. The insights that we receive are related to whether a process is a good candidate for RPA. For example, if it takes the human a lot of time to complete, or they are having trouble with it, then it might be suitable for RPA because putting a bot in place can optimize performance.
Another reason this is important is that human operators work very hard with day-to-day tasks, and they don't have much time to stop and look for processes that can be automated. Using task and process mining, it starts pulling out those insights. For example, it looks for the number of screens that the human is accessing and clicking on. It looks at each click, as well as every navigation and extraction. In the end, it generates a report for us.
The Agent Console has helped to decrease the average agent handling time, which is our main goal when it comes to these massive business operations. Average agent handling time is the metric that we primarily work with and as such, everything we do is related to reducing it. RPA in our use case is not used only to reduce the HC or FTEs; but it is used to boost this particular KPI too. In one of our use cases, we have had an average decrease of 30% in agent handling time, which is very considerable.
What needs improvement?
There are some enhancements that can be made within Orchestrator, such as the addition of new dashboards that provide us insights into processes that are already running, which would help us a lot.
For how long have I used the solution?
We have been using the UiPath platform since last year, at the beginning of 2020.
What do I think about the stability of the solution?
The stability of UiPath is very high, and this is a very important point. In fact, stability was the problem that we had when we were assessing the competition with UiPath. We evaluated several RPA tools and moved on with the UiPath partnership because it was the most robust. It is important to remember that we have a high availability environment, and the entirety of it must be stable. Our team is tool agnostic and extremely skilled in the largest RPA tool providers, enabling us to develop automation on any platform.
What do I think about the scalability of the solution?
UiPath is very easy to scale because, in every part of our solution, we use the Apps feature. This accounts for high availability and automatically provides us with scalability. For example, if we have a process that is handling 100 invoices at a time, but in three months we grow and need to instead process double that number, there is a feature that we can use to instruct the bots to run on additional machines. The scalability is very dynamic in this regard.
UiPath has a function within Orchestrator for dynamic allocation, where it can draw resources from a pool of machines in the infrastructure. For example, if there are 30 machines available then an option can be set to dynamically use the licenses. If there are 10 or 20 bots that need to perform tasks, the licenses will be automatically used to run the processes. You don't need to look at these machines to see what is running because all of the management will be taken care of automatically.
We currently have 25 staff involved in RPA. There are 20 developers and five architects, just to keep the projects and everything with the customers up and running. We expect this to improve and grow, doubling our numbers this year.
In fact, our expectation for growth is very high. Along with each implementation or development that we do, new opportunities arise. I would estimate that for each successful implementation that we have, four to five new opportunities are presented. Naturally, we will need to have more licenses and more contacts to increase the total number of bots in our environment.
From end to end we have about 50 robots already developed and running in a productive environment. We have human operators, back-office analysts, supervisors and coordinators involved in the whole process of execution and monitoring.
How are customer service and technical support?
After our initial deployment, we have rarely needed to be in contact with support. Our in-house team can do most things autonomously.
We have a premium support package from UiPath and they are very useful and very helpful. They help us with whatever we want and without any doubt support is one of UiPath's strengths. This is not just in a technical sense, but in terms of business and strategy, as well.
Which solution did I use previously and why did I switch?
As mentioned earlier, our RPA team is agnostic to the tool to be used. We can give our recommendation based on the customer's needs but if he has any premise that a specific platform should be used, we follow his needs.
How was the initial setup?
The initial setup is very straightforward. We have a very good team of architects and we also had some assistance from UiPath. It was not simple, but it was not too hard, either. It was okay and we didn't have any problems with the implementation. In total, it took between three and four weeks to complete the deployment.
In terms of strategy, I think for the best implementation, you need to first have the infrastructure set up. The infrastructure and architecture should be very well defined with UiPath because you have a bunch of functionalities that may or may not be useful, depending on the type of business.
Deciding what functionality is required is the very first step. Then, the second step is to have a methodology and a center of excellence for RPA, including frameworks and best practices. This will help to ensure that everything is implemented correctly and that you don't have problems in the future. Finally, you need to have certified developers and certified architectures because this is the most relevant part. You want bots to go live with the best quality to ensure customer satisfaction.
We now have the ability to provide this type of environment to a customer very quickly. We can configure the environment in between two and four hours, to have it fully up and running, and it is very simple to do. This is because we have an RPA infrastructure already built, so you just need to acquire the hardware that includes the machines and servers. Once they are up and running, we activate this section and we can develop and build the bots.
What about the implementation team?
UiPath was a big help during our initial setup. We have premium support, and they helped us with parts of the architecture, the infrastructure related to servers and the cloud, and getting it all set up properly in our high availability environment. It was like a four-handed job and it was all done well.
What was our ROI?
We have many different projects and customers and I would estimate that it has saved us and our customers something between $4M to 5M in total.
Just looking for our success case of the invoices processing, we avoid a cost of $10M.
What's my experience with pricing, setup cost, and licensing?
The price of UiPath is higher than competitors, although the cost depends on what functionality and tools you require. For us, we don't need anything extra in terms of functionality but our contract includes an extra charge for premium support.
Which other solutions did I evaluate?
We evaluated UiPath, Automation Anywhere and Blue Prism, all the best RPA tools providers in the international Market. As we are agnostic with the platform, we can develop using any of those three tools.
One of the cons of UiPath is the price. It's a bit higher than the other RPA tools. In terms of the pros for UiPath, it is more stable, it works in our high availability environment, the support is good, it is very agile and we can develop automations very quickly. Also, implementation was very fast and scalability is important.
What other advice do I have?
We have conducted a proof of concept using UiPath's unattended robot capability to enable a self-service chatbot. Specifically, we used the unattended bot to speak with our S3ND (messaging) solution, which is a chatbot. The scheduling was done via APIs and the communication was done through the chatbot. This is something that we have tested, but not yet deployed. We do think that this is an important next step for us to look at.
We do not use the Document Understanding and AI Fabric features at this time but we are already testing it in some of our customers to implement them as soon as possible.
Another feature that we do not yet use, but we are looking into, is using the AI Center to drag-and-drop machine learning models into RPA workflows. We recently had some discussions with the technical specialists at UiPath about the newer features that are available, and we are looking into arranging for training and webinars that will teach us how to use these new features correctly. Once we have a better understanding of how to implement them, we will begin looking for specific use cases.
The biggest lesson that I have learned from UiPath has to do with our customers and their operations. The most difficult and important challenge that we have is changing the mindset of our customers such that it is in line with digital transformation, and this is something that UiPath helps us with. They provide us with everything we need in terms of security, implementation, and high availability. Really trusting that these bots are doing the right thing is the biggest advantage that UiPath provides for us.
There are three main points that I would make for anybody who is considering UiPath. The first is the cost, in terms of money. The price of investment is high but the benefits are uncountable. Next, it requires that you look into what you really need, and whether it is all of the features that UiPath offers or just some of them. It is also very important that you look at your infrastructure because it has to be able to handle all of the bots. For example, we know that our processes need a lot of computing power and a lot of memory, so the hardware is important. This has to be built in advance of purchasing the software. The final part is the team, including the developers and architects. It is very important that they all be certified by UiPath. There is training and courses available, they make everything clear, and it includes learning the best practices, frameworks, and models to assure that you're doing everything right. If the company is audited and you are doing it properly then you won't have any headaches.
I would rate this solution a nine out of ten.
Which deployment model are you using for this solution?
If public cloud, private cloud, or hybrid cloud, which cloud provider do you use?