Scaling without losing control: a RedCore expert's view on automation

Yevhenii Yankovyi, Deputy CTO / VP of Technology at RedCore, shares his perspective on which processes should be delegated to systems, where humans remain irreplaceable, and why automation is a tool but not the ultimate goal.

Automation has stopped being a question of “whether to implement it or not” a long time ago. Today, it’s a matter of priorities and speed. Either you implement it and move ahead, or you keep hesitating and inevitably fall behind in the race for market leadership. At RedCore, automation is one of our key priorities. Based on personal experience, I’d like to explain when it’s time to introduce automation, where the risks lie, what should remain in human hands, and how to understand when your system isn’t working.

When the time comes

Automation appears in a business at a very specific moment when manual processes begin slowing growth down instead of supporting it. Up to that point, the team manages: tasks are distributed, agreements rely on people, and knowledge is passed from hand to hand. But once the business grows two or three times larger, this the workflow starts slowing down. People can’t keep up, mistakes multiply, and speed decreases.

People don’t scale

In digital businesses, I constantly see the same problem: key processes rely on specific people: tasks, decisions, operational actions. The issue is that human resources scale poorly. As a business grows, the amount of work increases and the workload on people rises proportionally. Eventually, one of two things happens: either business growth slows down, or the quality of work starts to decline.

Automation removes this limitation. Routine processes can be delegated to systems so that their volume can grow without proportional team expansion. Automation allows people to focus on tasks where they are genuinely more effective than machines: creative work, creative decisions, and handling exceptions. Everything repetitive that follows the same algorithm day after day can and should be delegated to systems.

According to Salesforce data for 2024, IT departments report the highest ROI from process automation at 52%, while operational teams report 47%. Workflow automation reduces errors by up to 70%, freeing people from monotonous tasks with a high risk of human error.

Increased control, not reduced control

Many people fear that automation means losing control. In reality, it’s the opposite. When processes are formalized and automated, you gain visibility into the entire picture: metrics, bottlenecks, and deviations.

The challenge lies elsewhere: in implementing automation correctly and managing changes within processes afterward. Processes constantly evolve, and the system must evolve with them. Flexibility is critical here. Humans adapt faster than systems: they notice a new problem and react, even when facing it for the first time. That’s why the question isn’t whether to automate, but what exactly should be automated.

Where humans are irreplaceable

The first thing to automate is a routine such as repetitive tasks performed according to the same algorithm, where decisions are predictable. The more often a task repeats and the less variability it contains, the higher its automation priority.

Humans are needed where there is no algorithm. Negotiations, force majeure situations, creativity, and everything that requires improvisation. Machines operate according to rules. Where rules do not yet exist, humans are indispensable.

Standards emerge through motion

Automation often runs into data quality issues, such as disconnected systems and incomplete data. Data inconsistency is indeed a serious challenge faced by nearly everyone. But in practice, the opposite mechanism works: as automation is implemented, standards begin spreading organically. The more systems operate using unified protocols, the wider these standards become adopted because compatibility becomes objectively necessary.

The best place to start is by defining interfaces and process touchpoints where data enters the system, where it exits, and how it transforms. This creates boundaries and constraints that guide processes in the right direction. These constraints are what provide predictability.

Companies use an average of 897 applications, but only 29% of them are integrated. According to Informatica, building a data pipeline can take up to 12 weeks, which is a critical delay for companies aiming to scale. The problem is not the absence of integration tools; they exist. The problem lies in the approach and the willingness to build processes systematically.

Team time is the only metric

How do you know automation is working? At RedCore, we believe the main indicator is the speed, quality, and amount of human involvement required with the same input data. How much time do people spend completing tasks? How many manual actions are needed? What percentage of errors occurs? These metrics should improve. People are the most expensive and the most valuable resource in any business. Therefore, the main criterion for automation efficiency is saving the team’s time and energy. If, after automation, people still spend the same amount of time, or even more, something has gone wrong.

McKinsey notes a dramatic shift in automation business cases: historically, ROI was calculated over 5–7 years, while today flexible robotic solutions and digital twins can reduce ROI periods to 1–3 years. When a project pays off within a year, companies can start small, prove value, and scale further.

Automation never ends.

How long does the transition take? There’s no instant result: simple tasks can be solved quickly, but full automation requires time. Technologies evolve, processes change, and every day brings new challenges and opportunities. What seemed 100% automated yesterday can be improved by another 20% today. Automation is a tool, not a goal. You can’t say: “We automated everything, the project is complete.” The more you automate, the more new tasks you discover that can also be optimized. It’s a continuous improvement process, not a one-time project with a fixed completion date.

Automate pain, not complexity

From a technological perspective, almost anything is possible. The tools exist, and the solutions are well established. The real question is the cost of implementation and scaling; sometimes, automation costs more than the value it creates, and that’s perfectly normal. But the greatest challenge is not technology, but maintaining flexibility. Industries are changing at tremendous speed: new requirements, new regulations, new business models. Systems must adapt just as quickly. Humans do this naturally: they see changes, assess situations, and make decisions. Machines still operate based on predefined algorithms, and reconfiguring them quickly remains difficult.

My advice: don’t start with the most complex process. Find something repetitive that frustrates your team every single day. Automate it within a month. If people genuinely feel relief, then you’re moving in the right direction. If not, then either you chose the wrong process or designed it incorrectly.

Automation is not about replacing people with machines. It’s about giving people back control over their time. Without it, scaling will eventually hit a ceiling. Sooner or later.