Switching vendors:
Success story about the transition from UiPath to Python RPA
BI Group is the most significant innovative construction holding in Central Asia. Using the Python RPA platform, the company saves its employees not only hundreds of hours of routine work but also a budget for IT products. We tell you how it succeeds in the material.
The company BI Innovations is engaged in developing IT in the BI Group. Finding and effectively applying the best IT tools is their key task. They started implementing RPA* in 2020. They began with simple processes in accounting, and over time, robotization covered most departments. For Forbes in Kazakhstan, the BI Innovations team shared the story of the transition from UiPath to Python RPA.
*RPA (Robotic process automation) is a technology for automating routine business tasks. Instead of humans, software robots perform tasks quickly and without errors, saving real employees' time. RPA can automate any routine tasks, such as report creation, data reconciliation, information search, and data digitization.
RPA-bots, instead of humans, work with any system through an interface. This integration is not necessary. Robotization is an effective tool on the way to digitalization with quick, visible effects.
They started building robotization on the UiPath RPA platform, which is a global leader in RPA. They gathered a team of developers, purchased licenses, and started writing software robots.
Finding an alternative to UiPath and transitioning
In early 2023, IT company leaders decided to change their approach to working with the RPA platform. One reason for the transition was the cost of the current software product. In a short period of time, it was necessary to find a new RPA product that would cover all business needs.
“We started by analyzing the RPA-platform market from among the most represented in Kazakhstan. Short-listed were two platforms, namely Python RPA and a product of the world’s largest software vendor. Then we started the selection process, called “system selection,” explained Arthur Bekmukhanbetov, IT Project Manager.
During the selection of the RPA product, the attention was paid to the following criteria:
- Functionality;
- Stability of work of robots and Orchestrator;
- High speed of migration to a new platform;
- Costs of licenses.
It was essential to move quickly to a new platform. For this purpose, the architecture and logic of robot development had to be similar to the previous platform. It was necessary to be capable of combining Low-code mode and programming. Moreover, the prior experience showed that it was important to rely on something other than expensive, narrow-profile developers, who are scarce in the labor market in Central Asia.
Also noteworthy was the cost of the license, as salaries in Central Asia are many times lower than in the USA and Western Europe, which directly affects the feasibility of using robots. Regarding the above criteria, the offer of the Python RPA platform from Kazakhstan matched BI Innovations. The cost of licenses was three times lower, and the quality was at the same level as the previous product. Additionally, Python RPA already had extensive experience implementing robotisation in large companies. Still, it was the first time our team faced the task of transferring from another platform.
Mukhamejan Duzenov, co-founder of Python RPA, recalls how complicated it was to convince BI Innovations to switch to Python RPA:
"We offered our product as an alternative to the product — one of the big solutions on the RPA market — used by BI Innovations. We were aware that it's hard to change a platform for robotization. Furthermore, we had to prove that our product is as good as our product and gives more development opportunities." Over the course of three months, the platform’s developers fine-tuned the product to meet the potential customer’s requirements. After a free personalized training course, the BI Innovations team quickly made several robots on their own on a trial version of the product — everything worked stably.
"BI Innovations believed in our product and allowed us to refine it. We improved the platform step by step — it was important for us to make a high-quality RPA tool. After three months of hard work, we got the result - our product fully met the requirements of BI Innovations," – said Mukhamedzhan Duzenov. "The cost of Python RPA licenses turned out to be the most optimal, namely three times lower than the previous platform. Python is the most popular programming language, so there were no problems finding developers initially; it will not be so in the future. We rewrote the previously developed robots to the new platform in just six months. The cost of ownership of the RPA platform became much lower. We can now robotise even small processes. The advantages of the new platform contributed to the acceleration of development four times," mentioned Kairzhan Bektembayev, Head of IT Department. From a technical point of view, BI Innovations is also satisfied with the product:
"Some processes can take even robots a very long time to complete. There was a task that took one robot five days to complete. We solved this problem by building the process architecture in Orchestrator. We divided an extensive process into five parts, and now one RPA bot works in parallel on five machines, on each of the five parts separately. Now, it takes one day to complete a task instead of five — shared Kasymzhan Asimov, RPA developer at BI Innovations. Switching vendors results and plans
BI Innovations is actively writing new robots on the Python RPA platform. The robotization results are attracting more and more holding company employees, and the robotization team is getting more tasks to automate with software robots.
"Working with a startup, in opposition to a large vendor, proved to be more convenient because of the team's flexibility. The so-called personal touch also played a role. Our team is actively "robotising" the Holding's internal business processes. We plan to introduce software robots in all structural divisions," said Kairzhan Bektem," shared with us Kairzhan Bektembayev, Head of IT BI Innovations. RPA is developing and gaining tools, such as AI and OCR, which are global trends. Right now, BI Innovations has its own OCR product in development. It will be used to solve the problems of recognizing paper documents. BI Innovations is also introducing machine learning to develop a software robot. The robot downloads bank statements from the Bank-Client application using business rules for the most frequent recurring payments, and a trained ML model categorizes the payments and enters them into the accounting system. The algorithm distributes the payments in a way that treasurers have done before.
From a technical point of view, the team is satisfied with the work of the Kazakhstani platform. Not only is it as good as the previous platform in terms of functionality, but it also gives developers more options:
"As a developer, I see the strength of Python RPA in the synergy of Low-code mode and the Python language power. For routine processes, it is convenient to use Low-Code. However, some processes are resource-intensive and need to be optimized. In these cases, you can use pure Python. It gives us great opportunities to work with new encryption methods and cryptography and helps us work with banks. Also, the optimization has freed up server capacity - the previous platform required seven servers, and now 5," summed up Kasymzhan Asimov, RPA developer of BI Innovations.
"Working with such a large customer is a great responsibility and reaching a new level of quality. Our team received feedback in real-time and finalized the functionality. They believed in our product and allowed us to work out the flaws. We are proud of the results and happy to receive positive feedback from users of our platform," summarized Mukhamedzhan Duzenov.