"A robot accomplishes tasks in one month equivalent to a year's effort by three humans." A startup from Kazakhstan automates routine business tasks

Python RPA is the sole entity in Central Asia crafting platforms for the automation of routine business tasks. The firm creates robot assistants for BI Group, Otbasy Bank, Air Astana, and RG Brands. These assistants ensure the accurate and timely execution of repetitive accounting and HR tasks, eliminating the risks associated with human error or fatigue. Consequently, businesses are safeguarded against potential fines for unregistered contracts or unpaid taxes.

In a joint project between Digital Business and Astana Hub called “100 Startup Stories of Kazakhstan,” the company’s founder and CBDO, Mukhamedzhan Duzenov, shared his entrepreneurial journey, starting from selling hay as a child to working in IT consulting and eventually founding his own startup focused on business automation. We also discussed how robots can automate tasks and ease the workload for employees in corporations, as well as the challenges that B2B businesses face in Kazakhstan.

“During my childhood, I gathered scrap iron, cut grass for hay, and sold newspapers and eggs”

— What sparked your interest in entrepreneurship?

— I began working during my childhood while growing up in a village near Kokshetau. The concept of business wasn’t prevalent; I simply needed money. I engaged in collecting scrap iron, cutting grass for hay, and selling newspapers and eggs in the city. Simultaneously, I pursued distance learning at a school for gifted children, focusing on exact sciences in which I excelled.

In the 10th grade, I enrolled in RPMS in Almaty. Later, through the Bolashak program, I pursued programming studies in Malaysia. My goal was to explore the world and unleash my potential. During an internship in Germany, I contributed to the development of a program for deep data analysis called RapidMiner.

In the last years of university, I came to the realization that coding wasn’t my forte. It demanded a level of diligence that, as it turned out, I lacked.

Upon my return to Almaty in 2011, friends recommended that I apply to one of the Big4 consulting companies. While facing rejections from KPMG and Deloitte, I secured a position at Ernst & Young. It was there that I crossed paths with Berik Tursynbek, eventually co-founding Python RPA. Our focus was advising companies on corporate IT, allowing us to delve into the establishment of IT processes within Kazakhstani companies and gain insights into their business needs.

— Did the idea of launching your own business begin to emerge during that period?

– Not immediately. Berik pursued studies at Carnegie Mellon University, and during that time, I took a temporary hiatus from consulting to work at the state company Zerde. Upon reuniting, we made the joint decision to embark on our own projects. My background included experience in management and consulting, while Berik brought expertise in software development to the table.

I began contemplating the technology that could set us apart from competitors in the market. It became apparent that in the USA and Western Europe, Robotic Process Automation (RPA) technologies were extensively employed. RPA involves the use of robots or bots to execute tasks for humans and facilitate the integration of multiple systems without relying on APIs. At that time, RPA was relatively unknown in Kazakhstan. Recognizing this as a promising opportunity to start an IT business, I decided to specialize in this technology and become an expert in the field.

— Can you explain what this technology entails?

— It functions as a versatile automation tool, creating mini-programs known as robots, without the need for coding expertise. These robots are designed to carry out tasks on behalf of humans, such as opening pages, downloading, and sending files. The process mimics human computer interaction, with the robot clicking buttons and moving the cursor, essentially performing tasks that would typically be done by an employee.

— How did you begin executing the startup concept?

— Fortuitously, we discovered that the pharmaceutical company Gedeon Richter was in need of assistant robots. After approaching them and presenting the technology, we successfully secured a contract. This led to a partnership with the US company, and we commenced selling and implementing the software.

It’s worth noting that in 2019, we were the sole company in the Kazakhstani market with a grasp of RPA technology. While there were clients, their number was limited.

— What factors contributed to the lack of popularity of the technology among companies in Kazakhstan?

—The product’s value was not well understood, necessitating us to function as RPA evangelists. We extensively shared information about the technology and its potential to enhance company efficiency. Essentially, we served as a marketing team for the US vendor, which did not actively invest in promoting the technology. Consequently, awareness about it was limited.

Another challenge was the high cost. RPA was beyond the financial reach of average companies. In smaller entities, identifying processes that would substantially impact labor costs through automation proved difficult. Implementing RPA wouldn’t be cost-effective for businesses with a headcount of fewer than 1,500 people.

In 2019, there were no affordable alternatives available on the market. This led us to conceive the idea of creating more budget-friendly software for Kazakhstan companies. We identified an open-source solution and reached an agreement with the author to use his development as the foundation for our own solution. This marked the beginning of the development of our platform, giving rise to Python RPA. Prior to this, we operated under the name Seiko Lab.

“Clients can create robots independently without the need for extensive programming experience”

— As you began selling your proprietary solution, who were your competitors in the market?

— We were the sole producers of such software in Central Asia, with US and Russian competitors in the market.

— What distinguishes your offering from theirs?

Primarily, our emphasis is on speed. Our platform enables the straightforward and rapid creation of robots, requiring minimal programming expertise.

Secondly, our solution is notably more cost-effective. Robots yield returns within 6-9 months, whereas foreign counterparts take three times longer to pay off.

Thirdly, the platform utilizes the Python language, enabling clients to create their own robots using clean Python and manage them through our server. While there are platforms solely focused on orchestrating Python scripts and others on providing a no-code/low-code environment for robot creation, we encompass both aspects.

Most importantly, we bring extensive experience to the table. With over 200 cases of how to integrate robots into business, there’s no need to invest time in searching for processes. Clients can easily select them as if choosing from a menu.

— Do you primarily sell the platform or specialize in developing robots for companies?

— Over the past six months, we have been progressively transitioning from robot development to becoming a comprehensive software producer. Robot development is a labor-intensive process, and to avoid direct implementation, we instituted a partner program. This involves clients and partner integrators who independently undertake the task of writing robots.

— What functions can the robot perform?

— The solution finds application in accounting, procurement, IT, and treasury. We develop a virtual assistant for employees capable of executing repetitive actions seamlessly, without pauses.
Here’s an accountancy example: quarterly, there’s a need to reconcile tax data in the internal accounting system with information on the state portal of electronic tax invoices. Two Excel files are downloaded and reconciled, a process that takes time. Humans are susceptible to errors, can fall ill, or may forget. The robot automates this task through a predefined algorithm: it downloads the files and presents the accountant with the results, including any differences, if present.

In Kazakhstan, employees are mandated to submit data regarding the initiation and termination of employment agreements on enbek.kz within 3-5 days. HR personnel traditionally handle this task manually, which introduces the risk of errors or missing deadlines. This challenge becomes more pronounced with a large headcount. The robot assistant is designed to autonomously register agreements without human intervention.
Robots have the capability to associate electronic invoices with documents confirming the receipt of goods or services, generate reports on receivables, and verify the reliability of counterparts or debtors.

Upon closer examination, the essence lies not only in time savings. We alleviate employees from monotonous tasks to prevent burnout. Additionally, we minimize the risk of errors, as robots are not prone to mistakes like humans.

— Could you offer instances of how companies employ robots?

— In a major railway company, the automation of the employment agreement registration process took place. The task involved registering a substantial number of old agreements on enbek.kz. An estimation indicated that it would take three individuals one year to complete the registration manually. However, a robot accomplished the task within one month.

In a prominent FMCG company, the process was automated to streamline the work of marketologists with reports in the Nielsen system. Robots are responsible for downloading data for analytics from advertising accounts.

At Optima Bank, developers underwent training to autonomously create robots, and within six months, they automated 40 processes associated with reconciling settlements with payment systems, transit accounts, and partner banks. This represents a significant accomplishment.

— Given the extensive automation implemented, do companies reduce their workforce when robots undertake their tasks?

— No, at most, some companies eliminate job openings. We don’t advocate for layoffs. From a karmic perspective, I believe it’s wrong. Moreover, our clients do not terminate employees, as the robot saves between 30 minutes to 2 hours per day. In rare instances, individuals may spend the entire day on routine tasks.

“There’s Mercedes, and then there’s Hyundai Sonata. Thus, we are akin to Hyundai Sonata”

— Who constitutes your main clientele?

— Our primary clients are companies with a workforce of 500 people or more. The industry itself is not a critical factor, given that the Kazakhstan market isn’t expansive enough to exclusively target a single sector.
Most frequently, we deploy RPA solutions in banks, addressing numerous small processes associated with reconciliations. Additionally, our services extend to railway companies. Our clientele includes the construction company BI Group, the cloud infrastructure operator QazCloud, Kostanay Minerals, and the waste disposal company Tartyp.

— How difficult is it to sell an RPA solution?

— We typically highlight that our product is comparable to US counterparts, with a notable advantage of a significantly lower price. While there might be some functionality differences, the cost advantage is substantial. Drawing a parallel, it’s like comparing Mercedes to Hyundai Sonata, and in our case, we align with Hyundai Sonata. Most reasonable clients respond with, “Alright, let’s proceed.” We’ve developed proficiency in replacing foreign software.

Our sales cycles typically span from 6 to 9 months, and the primary obstacle we face is rooted in psychology. In our region, there is a lack of appreciation for locally developed high-quality software. There exists a preconceived notion that our capabilities are limited to mining mineral resources, leading to the perception that we are not an IT-driven nation merely because we are Kazakhstan.

— In your opinion, how can this be changed?

–It would be beneficial if our government were to declare: “Here’s a company with a product no less competitive than others; let’s provide them with support.” This would quickly attract numerous clients. The software used by the oil mining company is irrelevant; the volume of oil remains unaffected by this factor. If the government extends support to all local startups in this manner, it would significantly simplify the process of commercializing projects on the global market.

Currently, there are no SaaS companies in Kazakhstan’s B2B sector with an estimated value exceeding $100 million that operate in foreign markets. Building such an ecosystem necessitates the presence of entrepreneurs capable of creating unicorn startups, a thriving retail market, scientific centers, and large companies willing to purchase emerging businesses. However, Kazakhstan currently lacks this ecosystem. Nevertheless, with time, determination, and motivation, this issue can be resolved, with the crucial involvement of the state.

We appreciate the assistance we have received, including a grant of 30 million tenge from QazInnovations. However, when compared to countries like Malaysia, where startup grants start from $500 thousand, this amount seems comparatively small.

— Have you achieved any operational profit so far?

— To be honest, we haven’t conducted any specific calculations. It is possible that we have already achieved some profit, although it is certainly not substantial. All of our earnings are currently being reinvested into our products. We have successfully completed two funding rounds, namely pre-seed and seed, with a total investment in the project amounting to no less than $350 thousand.

— How do you plan to increase your earnings?

– To become a successful B2B company, you have the option of pursuing two paths. The first involves securing substantial funding rounds to invest in marketing and promotion, enhancing the product’s quality and visibility.
The alternative approach involves moving consistently along new, undiscovered trails. We have opted for this path, aiming to deliver value for reasonable compensation. Our current strategy involves entering foreign markets, where there is potential for increased earnings.

“Without venturing into external markets, we would either die or remain a small local company”

— What is the level of your international recognition?

– We conducted software sales trials targeting small and medium businesses in foreign markets such as Germany, the USA, Indonesia, and the Netherlands. We have accumulated over 500 purchases, with our clientele comprising freelancers, IT agencies, and small companies.

We ventured into the SaaS solutions market in the US to obtain valuable feedback and enhance our brand recognition. Our efforts have garnered some popularity. Currently, we are in discussions with the manufacturer of silicone implants, Silimed, and have already engaged in conversations with the marketing firm Nielsen and the electronic equipment supplier, Schneider Electric.

— How do you establish relationships with customers from abroad?

— B2B sales heavily rely on networking there. It is not possible to simply approach a company from the Fortune 500 list (which comprises the largest US companies in terms of revenue according to Fortune magazine). These companies often have long standing supplier relationships spanning 30 years. Our strategy involves working through partners, and we have already established partnerships in Kenya, Nigeria, UK, and India.

— What are your current plans?

– We aim to fortify our presence in the local market and secure positive cash flow before venturing into international endeavors. Following the New Year, I intend to take proactive steps. I am applying for foreign acceleration programs and searching for an individual who will assume responsibility for sales in Kazakhstan on my behalf.

Our global objective is to evolve into an international provider of RPA solutions—there’s no alternative. Restricting our development solely to Kazakhstan would result in either death or the fate of a small local company. Similarly, from a financial perspective, we could have pursued selling doner kebabs, but our aspirations extend beyond that. We are pursuing our big dream.

Come what may. If we succeed, it will be fantastic. If not, it will serve as a valuable lesson for other startups. We’ll openly share our mistakes, providing useful insights for those planning to embark on a similar journey. That’s how the startup ecosystem operates.