The road of change
Kairat Bazarbekov, CIO at Centras Group, Kazakhstan, talks about his automation experience at Centras Group: how they gradually managed to combine several disparate IT systems into one and what approaches help in business automation.
In an interview:
- The era of ERP systems: why and how applications of this type were implemented.
- Centralization of IT services: from several IT departments with separate server rooms to one corporate IT hub and unified infrastructure.
- Change in planning period and methodology: from waterfall to Ajail, from annual plans to 2 - 3 month projects.
- Changing technology priorities: moving to web development and mobile applications.
- Solving the staffing problem: unexpected approach, downsizing the team.
- How we learn new things: conferences as a way to quickly understand new trends.
- Current priorities: decomposing ERP systems and automating development.
- Tips for the IT leader
The interview is recorded in the speaker's language, we suggest you to use subtitles.
The era of ERP systems
Centras Group has been operating since 2004. There are four companies in it: two insurance companies, one medical company, one investment company, totaling about 1,500 people. These businesses grew gradually, starting with the investment business. In the beginning, IT in all the companies developed independently.
In 2017, I took up my current position after leading the operations department. The main task at that time was to automate internal processes by implementing an ERP system. The main reason for implementation was the requirements of the National Bank of Kazakhstan on transparency of reporting.
After many mistakes along the way, we learned to be productive. The key word and the key to success is "elephant carpaccio". Having developed the team, we implemented a lot of ERP functionality and realized what we lacked competencies to implement.
We used different products - both Western ERP systems and Bulgarian applications, and in the insurance company we implemented the package of the company "Insurance Laboratory", our long-term partner. Their solution is used in two of our insurance companies. It is architecturally correct, scales well and has proven to be reliable.
Centralization of IT services
Just at that time there was an active discussion in the group of companies about how to change the IT department so that it could keep up with the business needs in automation. The idea of uniting IT resources of the whole group, not only IT, but also other back-office divisions, so that each such division would serve all the companies of the group. We came to this already after the pandemic, which showed the power of remote working.
In 2020, we united the previously disparate IT services of the companies into a single structure for the entire Group. It is this structure that implements the tasks set by the business. It includes all IT specialists: business analysts, devops, and mainly, of course, developers.
We have gone a long way towards unification and standardization of IT services in the group of companies, including infrastructure. This has allowed us to gradually move towards proactive management of networks and computing resources. On the other hand, this allows business analysts to offer clients existing solutions rather than trying to develop something new in response to every business desire.
Changing the planning period and methodology
Six years ago, we realized that planning the budget for the year was hindering our development, so we switched to semi-annual budgets. Now we do KPI planning on a quarterly basis. Not only the business wants to see quick wins in short steps, but also IT. We do not have KPIs coming down from above. Department directors set their own goals and suggest ways to achieve them, including the introduction of IT tools. Major KPIs are defined at the company level, and smaller ones are prioritized, starting with the requirements of regulators. This takes place at ICOM, the board of top managers.
The taste for automation was instilled by the introduction of the first ERP system. Even then it became clear to business executives what real effects the right automation could produce, and the appetite grew rapidly. IT projects are prioritized for 2 -3 months, and individual tasks are planned for this period, but taking into account further steps. Large projects are decomposed.
After 2 -3 months the business gets something that works and gives results. If it turns out that this is exactly what was needed, the work continues. If not, the project is canceled. We used to use the traditional waterfall for periods of six months or more. Changing the planning period and moving to agile methodologies is clearly yielding positive results.
Communication used to be very difficult, and sometimes department directors would try to explain directly to developers what exactly they wanted. But they explained it in business terms, so they were usually unsuccessful. Now my goal is proper communication with the business, the product owners - that's what's important. Otherwise we keep getting ahead and they lose touch with us.
I think any company goes through such a stage in its development. We have had many conferences over the last year, and at them it becomes clear that the situation is about the same for many people, regardless of the industry.
Changing technology priorities
We had a lot going on in our ERP system, but the pandemic and the time since has shown that needs change quickly and we need to go web-based. The back office can work in the office or connect via vpn remotely. But all salespeople need to be able to close deals directly from their phone and go through all stages of the contract painlessly. For this we need web technologies.
We did not have enough developers of our own in this area, so we invited outsourcers. It was not always successful. Some of them fulfilled our tasks, but not completely, and then we had to take the development back to our company and finish everything that was not done.
On the one hand, it was the result of our own mistakes, not fully clearly formulated TOR. On the other hand, our end customers, business managers did not fully understand what the result should be. In such conditions it is better to develop the necessary expertise internally. The use of outsourcing and outstaffing depends very much on the business. We can use them, but we still have to take all the products in-house for support. That is why we mainly develop our own development.
Solving the staffing problem
We don't have any very large-scale tasks to outsource, and we take individual specialists in narrow areas from the market. There are almost no seniors, everyone is fighting for them, so we mainly try to grow our staff internally and cooperate with various courses and schools that train programmers.
Our staff grew to more than 80 people a year and a half ago, and then I realized that it was no longer realistic to manage such a staff in our situation. We were constantly hiring juniors, with three or four out of every ten remaining. But seniors and even middles are chronically lacking, I often had to manage the development myself. When it became clear that we were close to chaos, we stopped accepting new applications from the business for a month. We sorted out everything we had, prioritized, and reduced the number of projects and staff. We left 30-40 projects out of a notional hundred. Some tasks were simply poorly described.
At the same time, we restructured the teams, left no more than 7-8 people in each team, and increased the responsibility of team leaders. Now we have 53 people working in development. Of them, 30 are web developers, 10 are ERP developers, and the rest are mostly business analysts. We have reduced the number of people so much over the year. But the business also began to realize that it is impossible to cover the vastness. Gradually we moved to the concept of product owners, so that the business also understood its responsibility.We abandoned the concept of an "incubator" because so many novice developers left us after six months to a year. Of course, it is honorable to be a forge of personnel, but it is not very profitable.
The matter was complicated by the fact that one HR-department was looking for all the personnel for the whole group of companies, without taking into account the specifics of IT. When we recruit new people, it is not so much the hard-skills that are important, which can be pulled up relatively easily, but the general level of culture, the ability to ask questions, and the desire to develop.
A good resource is current employees of the company working in other divisions. Among them there are those who are interested in IT, who want to be involved in process automation. We are happy to transfer them to us and train them.
How we learn new things
A good example is the transition to DevOps. Previously, like everywhere else, we had system administrators engaged in infrastructure management, while development was separate. The functions of automating testing and new releases were partially performed by developers themselves, but there was no systematic approach. Wanting to understand this, we organized a DevOps Days forum. Our business leaders heard a lot of completely new, previously unfamiliar terms. But the more presentations they listened to, the longer we talked to them about this new approach, the more understanding there was. Now we have largely automated the output of new releases into production and continue to actively develop this area.
For us, forums are a proven way to understand an unfamiliar area. We have been participating in the Kazakhstan Grouse Forum for more than ten years. Among other things, we have already organized the Kazakhstan Technology Summit four times. We invite experts with real cases, including those from abroad. Organization starts six months in advance. Even at this stage, you begin to recruit expertise.
We have also had cases when we started to deal with unfamiliar technologies, and then it turned out that we didn't need them, that there was no point in dealing with them, primarily for economic reasons: there would be no return. We had to stop development, to wind down projects.
We introduced a general rule: until the business formulates a task, we only conduct testing. We start scaling up when there is clarity.
The main business processes of companies are automated in ERP systems, there are several of them. But to build BI and for many other things it is necessary to separate applications, reduce the load on them, and "glue" them differently than in ERP packages.
Another priority is further automation of development. It is closely related to the global task: to solve business tasks faster, to bring new applications to the market faster.
Tips for the IT leader
Change management in IT is important. Every innovation and every trend should be evaluated in terms of its usefulness for your company's business. It's impossible to chase all the innovations.
Communication is a key skill. On the one hand, there is the constant search for information in telegram channels, at meetups, and in various communities. On the other hand, communication between development and business. It's not so important to understand the technology in detail; there are tech leads and technical directors for that. It is fundamental to be able to assess risks and explain to the business what is possible, what is impossible, what are the potential problems.
Initiative and the ability to propose new processes to replace the old, familiar ones is important. Many people are afraid of automation, afraid of being fired. There is a big gap between business and IT. We need to build as many bridges across it as possible, so that the business understands how its tasks are solved. We also need visionaries, followers in the business environment. The main question is how to build business processes.
The third important component is the management of one's resources, their arrangement and interaction between them.
Artificial intelligence technologies have been developing for decades, but only in the last few years have neural networks "smartened up" to a level that allows them to be used as assistants or even to replace human labor.
The beginning of the year was marked by a flurry of activity by various legislators to deal with artificial intelligence. The New York Times filed a lawsuit against OpenAI and Microsoft, and the UAE created an AI council to accelerate the adoption of AI systems in critical sectors of the economy.
What should a business solution be, so that it is chosen by young people, but at the same time not rejected by older professionals? Maybe beauty will save the peace in the team, because everyone wants to work in an application that looks beautiful? It is true, but User Interface (UI) is good when it is based on User Experience (UX). If harmony between UI and UX is achieved, the product will be both beautiful and user-friendly.