Alexey Samoilov: 'The Responsibility Doesn't Belong to the Software'

The UK could face a shortage of more than 43,000 technicians qualified to work on electric vehicles by 2035, according to the Institute of the Motor Industry. As electric vehicles become more complex, workshops also need specialists who can interpret increasingly complicated diagnostic data and make sound repair decisions.

For Alexey Samoilov, an expert in high-voltage vehicle systems, that distinction is central to modern vehicle repair. With more than 20 years of industry experience, he holds the German professional qualification Kraftfahrzeugtechnikermeister — Master Automotive Technician — and has specialist certifications from Mercedes-Benz and BMW. His work focuses on complex electronic faults and high-voltage systems in electric and hybrid vehicles.

Samoilov has developed and applies his own structured approach to high-voltage diagnostics, combining manufacturer diagnostic systems with electrical measurements and signal analysis. In 2025, he received the Velto European Award in the Sustainable Futures Leadership category for his work in EV and high-voltage systems diagnostics. And in August he was invited to join the jury of the Global 1000 Award in the Sustainable Impact & Green Tech category.

Samoilov's central argument is straightforward: a diagnostic system can identify a problem without necessarily explaining what caused it.

A Fault Code Is a Starting Point, Not a Verdict

Modern electric vehicles rely on interconnected electronic control units, battery management systems, sensors and communication networks. When one part of that system behaves unexpectedly, several other components may report errors.

'A scan tool can collect codes in seconds, but interpreting them requires an understanding of how the systems interact. Diagnostic software doesn't repair cars, only people do,' Samoilov says.

The distinction matters because a fault code does not always point directly to a defective component. It may indicate a symptom, a communication failure or an abnormal operating condition rather than the underlying cause.

'An oscilloscope can show the electrical signal itself, while network analysis can help establish whether control units are communicating correctly. The technology has become much smarter over the years, but it still can't replace engineering judgement,' the expert explains.

This does not mean diagnostic software is unreliable or unnecessary, but findings need to be interpreted in context. The same warning can have several possible explanations, and the job of the technician is to determine which one fits the evidence.

The Price of Misdiagnosis

One case from Samoilov's workshop illustrates the difference between reading an error and understanding it. A modern vehicle repeatedly generated fault messages involving the high-voltage battery and powertrain electronics. Initial computer diagnostics pointed towards a possible failure of an expensive control unit or power electronics module.

But rather than relying on the fault codes alone, Samoilov examined the electrical signals using an oscilloscope. The measurements revealed irregularities in the CAN communication signal — the network through which electronic control units exchange information. The problem appeared under particular operating conditions, including vibration, making it difficult to reproduce during a routine diagnostic check.

Further investigation pointed to an electrical disturbance on the communication line or a poor connection at a plug. After the connection problem was repaired, the vehicle returned to normal operation, and the proposed replacement of the expensive component was avoided.

The significance of the case lies in the sequence of decisions. The diagnostic system identified a genuine problem, but its initial findings did not establish the root cause. Additional measurements allowed Samoilov to test the original hypothesis against an alternative explanation.

'Software gives you an interpretation, not necessarily the complete picture,' he states. 'For a workshop, that difference can determine whether a repair addresses the fault or merely replaces a component associated with it.'

How Diagnostic Errors Drive up EV Repair Costs

The economics of electric vehicle repair make this distinction particularly important. High-voltage batteries and power electronics can be expensive to replace, and a mistaken diagnosis can turn a relatively minor repair into a major bill.

In his interview responses, Samoilov illustrates the potential difference with two figures: a connector may cost only a few euros, while replacing a high-voltage battery can cost €15,000 or more. These figures describe the possible scale of the cost difference; they are not a measured saving from the communication-fault case.

'The financial risk is not limited to battery replacement. A workshop that misidentifies the source of a fault may face repeat visits, warranty claims and disputes with customers,' the expert remarks.

A correct diagnosis can require more time and specialist equipment at the outset, but that investment may be less costly than replacing a functioning component or carrying out the wrong repair.

For owners, this can mean avoiding an unnecessary major expense. For workshops, it can mean fewer repeat repairs and stronger customer confidence. For manufacturers, accurate fault identification can help distinguish component failures from problems elsewhere in a vehicle's systems.

A Structured Method for Complex High-Voltage Faults

Samoilov's diagnostic approach evolved through years of hands-on experience in the workshop. His methodological guide 'Diagnosing High-Voltage Faults: From the DTC to the Engineering Cause of Failure' sets out a structured process for investigating complex faults in electric and hybrid vehicles.

It begins with understanding the vehicle's high-voltage architecture and operating logic, then examines when and under what conditions the fault occurs. The next step is to localise the problem through appropriate, safety-conscious testing. Before recommending a component replacement, the technician must establish that the evidence supports the decision.

The method is designed to prevent a common diagnostic mistake: treating the first plausible explanation as the final answer.

'Before recommending the replacement of an expensive component, I want to be certain that the diagnosis is supported by evidence, not assumptions,' Samoilov says.

His approach is not a replacement for manufacturer procedures. It complements them by providing a framework for investigating cases in which the initial diagnostic information does not fully explain the fault.

Training Specialists To Think Beyond the Error Code

Alexey Samoilov is a member of the Institute of Electrical and Electronics Engineers, one of the world's leading professional organisations in electrical engineering, electronics and technology, connecting specialists internationally and contributing to the development of technical standards and knowledge across the industry. And he has spent years training automotive mechatronics technicians within Germany's dual vocational education system. He believes that technicians need to learn how to reason through unfamiliar faults, rather than memorise a list of codes and corresponding repairs.

'Collecting data isn't the same as understanding it,' he says. 'The real skill is recognising which values matter and how they're connected.'

Samoilov also participates in Mercedes-Benz technical forums, where specialists exchange experience on difficult cases and emerging technical issues. Discussions increasingly cover software-defined vehicles, high-voltage diagnostics, remote diagnostics and cybersecurity. The value of this exchange lies in keeping professional knowledge current as vehicle systems evolve.

In August Samoilov was invited as a jury member in the Sustainable Impact & Green Tech category in the Global 1000 Award. His 2025 Velto European Award recognised his work in electric vehicle and high-voltage systems diagnostics.

'AI can process huge amounts of diagnostic information in seconds but someone still has to decide whether that recommendation is justified, because the responsibility doesn't belong to the software,' Samoilov maintains.

For Samoilov, the future of vehicle diagnostics will involve more automation, more data and more sophisticated software. Yet the decisive moment will often remain the same: a specialist must decide whether the evidence is strong enough to support a repair.