Practical interview techniques I use to verify CV claims, detect inconsistencies, and avoid hiring based on inflated experience.

Most CVs don’t contain direct lies.

They contain something more subtle and harder to catch: carefully shaped reality.

Job titles get upgraded. Responsibilities expand slightly with each rewrite. English level becomes “fluent.” Side projects suddenly turn into “lead roles.” Nothing is obviously false, but the overall picture slowly drifts away from reality.

And the problem is not only candidates. It is the system we all use. A CV is not designed to verify truth. It is designed to present the best possible narrative.

That’s why, during interviews, I stopped treating CVs as truth. Instead, I treat them as a hypothesis.

Then I verify it.

Here is how I do it in practice during early-stage interviews.

Education: I ask for proof when it matters

Education is one of the easiest things to “smooth over” on a CV. Not necessarily fake, but often exaggerated, unclear, or incomplete.

Instead of debating it, I simply ask for confirmation when it is relevant to the role.

A diploma, certificate, or official document is usually enough.

This is not about distrust. It is about removing ambiguity early. If someone truly graduated, this is never a problem. If they didn’t, it becomes obvious very quickly.

English level: I switch languages immediately

“Fluent English” is one of the most common overstatements in recruitment.

The fastest way to validate it is also the simplest: I start speaking English.

A 5-minute conversation is enough. No formal test, no preparation, no warning.

What matters is not grammar perfection. I look for:

• comfort in spontaneous conversation
• ability to explain ideas clearly
• reaction when they don’t know a word
• whether they can keep the conversation going naturally

The gap between CV claims and real ability usually appears within the first few minutes.

Experience: I verify it at the source

Work experience is where CVs tend to drift the most.

Titles become inflated. Responsibilities expand. Timelines get adjusted slightly.

When experience is important for the role, I don’t rely on descriptions alone.

There are two reliable methods:

First, documentation. Employment contracts or confirmations can validate basic facts.

Second, reference checks. With permission from the candidate, I may contact previous employers to confirm role scope and responsibilities.

On top of that, I also use behavioral questions related to real situations that are likely to have occurred in that job. Instead of asking generic questions, I try to anchor the conversation in specific work scenarios — for example how a candidate handled a production incident, a difficult stakeholder, a missed deadline, or a system failure. The goal is to see whether their answers are consistent, realistic, and aligned with the actual level of responsibility they claim.

This is not about catching people off guard. It is about checking whether they have truly experienced what they describe, or whether they are reconstructing it from assumptions.

The goal is simple: confirm that the story matches reality.

Technical skills: I avoid relying only on online tests

Online coding tests are convenient, but they are not reliable indicators of real ability anymore.

They can be rehearsed, optimized for, or in some cases outsourced.

Instead, I focus on live technical evaluation.

A real-time technical discussion in the office or via screen sharing tells you far more than any automated test.

I look at:

• how the candidate approaches an unfamiliar problem
• whether they can break down complexity
• how they think under slight pressure
• whether they understand trade-offs, not just solutions

I am not looking for perfect answers. I am looking for real thinking.

Soft skills: mostly visible in the first conversation

Soft skills are often over-engineered in hiring processes.

In reality, most of them are visible immediately.

Communication style, clarity, confidence, and openness appear naturally in the first interview.

One additional signal I pay attention to is context outside of work. Interests, personal projects, or curiosity often reveal how someone thinks and whether they grow beyond task execution.

It is not a formal scoring method. It is context.

The real goal is not to “catch lies”

The goal is not to accuse candidates of lying.

Most candidates are not trying to deceive anyone. They are trying to present themselves in the best possible light, often under pressure to stand out in a competitive market.

The real issue is uncertainty.

CVs are incomplete signals. And when hiring decisions are based on incomplete signals, mistakes are inevitable.

So the goal of an interview process should be simple:

Reduce uncertainty early.

Not by being suspicious. But by validating the key claims that matter for the role.

Because in a world where CVs are becoming increasingly polished and AI-assisted, clarity is more valuable than ever.

W takich warunkach pojawiają się narzędzia do automatyzacji selekcji CV. Jednym z nich jest Hero Parser, który sprawdza dopasowanie kandydatów do oferty pracy na podstawie umiejętności, doświadczenia, języków i innych kryteriów.

Gdzie takie podejście ma sens, a gdzie może zawodzić?

Jak działa Hero Parser w praktyce

Na poziomie koncepcji narzędzie robi jedną prostą rzecz. Przetwarza CV i zamienia je na zestandaryzowane dane, które można porównać z wymaganiami oferty pracy.

W praktyce oznacza to:

• wyciągnięcie umiejętności z CV
• analizę doświadczenia zawodowego
• dopasowanie do wymagań oferty pracy
• przypisanie wyniku, który pozwala sortować kandydatów

Różnica względem klasycznych ATS-ów polega na tym, że nie kończy się na przechowywaniu CV. System faktycznie próbuje je interpretować.

Co sprawia, że rekrutacja staje się ustrukturyzowana

Jednym z najważniejszych elementów Hero Parser nie jest samo parsowanie CV, ale sposób, w jaki porządkuje cały proces oceny kandydatów.

W większości systemów ATS CV są jedynie zbierane i filtrowane po prostych kryteriach. W praktyce każdy rekruter i tak ocenia je inaczej, często intuicyjnie.

Tutaj podejście jest inne. Każdy kandydat jest oceniany według tych samych zasad i tych samych kryteriów. To sprawia, że porównywanie osób przestaje być chaotyczne i subiektywne.

System nie ogranicza się też do listy technologii, co jest typowe dla wielu ATS-ów. Umiejętności są tylko jednym z elementów. Dużo większy nacisk położony jest na doświadczenie:

• jakie projekty kandydat realnie realizował
• w jakim kontekście używał danych technologii
• jak długo i w jakiej roli pracował

To zmienia perspektywę, bo w praktyce to doświadczenie, a nie sama lista umiejętności, częściej decyduje o dopasowaniu do roli.

Dodatkowo system obejmuje także ocenę rozmowy wstępnej. Zamiast luźnych notatek rekrutera, pojawia się ustrukturyzowana ocena, którą można zestawić z wynikami analizy CV. Dzięki temu cały proces zaczyna być spójny, a nie składa się z niezależnych, subiektywnych opinii.

Feedback dla kandydatów jako część procesu

Jednym z najbardziej nietypowych elementów Hero Parser jest automatyczne generowanie feedbacku dla kandydatów.

I tutaj pojawia się ciekawy kontekst z rynku.

Jeden z rekruterów napisał:

To podejście zmienia jedną istotną rzecz. Nawet jeśli kandydat nie przechodzi dalej, nie zostaje z niczym. Dostaje informację:

• jak wypadł na tle innych kandydatów
• które wymagania spełnił, a których nie
• dlaczego został odrzucony

W praktyce to nadal rzadkość. W wielu firmach kandydat nie dostaje nawet automatycznej odpowiedzi o statusie.

Co działa dobrze

Największą wartością jest uporządkowanie procesu w sytuacji, gdy jest dużo kandydatów.

Przy dużej liczbie CV ręczna selekcja szybko przestaje być efektywna. W takich przypadkach ranking kandydatów pozwala szybciej dojść do sensownej shortlisty.

Dobrze działa też sama standaryzacja oceny. CV zaczynają być porównywalne, zamiast interpretowane indywidualnie przez każdą osobę w zespole.

Dla kogo to ma sens

Hero Parser ma największy sens w firmach, które:

• regularnie rekrutują
• dostają dużą liczbę CV na jedno ogłoszenie
• chcą skrócić etap wstępnej selekcji

W mniejszych procesach rekrutacyjnych, gdzie kandydatów jest kilku lub kilkunastu, korzyść jest dużo mniejsza.

Podsumowanie

Hero Parser nie rozwiązuje rekrutacji jako całości. Nie zastępuje decyzji rekrutera i nie „wybiera najlepszych ludzi”.

Zamiast tego robi coś bardziej praktycznego. Wprowadza strukturę tam, gdzie zwykle jest chaos.

Porządkuje CV, ujednolica sposób oceny i dodaje element, którego w rekrutacji brakuje najbardziej — feedback dla kandydatów.

I właśnie ten ostatni element wydaje się najbardziej niedoceniany.

In 2026, IT recruitment has hit a strange point.

Every CV looks good. Structured. Clean. Tailored to the job description.

And yet, choosing the right candidate is harder than ever.

Why all CVs look the same in 2026

The short answer: AI.

Most candidates today use AI tools to write or refine their CVs. This leads to consistent formatting, clearer descriptions, and better alignment with job requirements.

But here’s the important part:

AI doesn’t fake experience.

It organizes it.

Candidates still describe real projects, real technologies, and real responsibilities. The difference is that now everything is presented in a more professional way.

So when people say “AI is ruining recruitment”, they’re pointing at the wrong problem.

The real problem with AI CV screening

The real issue is how companies screen CVs.

Most systems still rely on keyword matching. They scan for terms like “Java”, “Spring”, “Docker”, and assign scores based on their presence.

This approach worked when CVs were inconsistent and incomplete.

It doesn’t work anymore.

Today, almost every CV includes the right keywords. Candidates know what to include. AI helps them structure it properly.

As a result, recruiters are left with dozens — sometimes hundreds — of “perfect matches”.

But they’re not actually equal.

Why keyword-based recruitment is failing

This is where most discussions around “AI in recruitment” go wrong.

The problem isn’t AI-generated CVs.

The problem is keyword-based evaluation.

In IT recruitment, two candidates can have identical tech stacks on paper:

• both know Java
• both worked with Spring
• both used Git

Yet their real value can be completely different.

One may have worked on small internal tools for a few months. The other may have spent years building large-scale production systems.

Keyword matching treats them the same.

That’s the failure.

How to evaluate candidates beyond keywords

If you’re searching for “how to screen CVs more effectively” or “how to assess real experience in IT recruitment”, the answer is relatively straightforward — but not easy to implement.

You need to shift from keyword detection to experience analysis.

That means looking at:

• how long a candidate used a given technology
• the scale of the projects
• the complexity of systems
• level of ownership and responsibility

All of this information already exists in CVs.

Traditional systems just don’t use it.

AI didn’t break recruitment — it exposed its weaknesses

It’s tempting to blame AI for making recruitment harder.

Some companies even ask candidates not to use AI when applying.

But that’s treating the symptom, not the cause.

AI didn’t create identical candidates.

It revealed that existing screening methods were too shallow.

When every CV becomes clear and well-structured, weak evaluation logic becomes obvious.

A better approach to AI CV screening

If you’re thinking about “AI CV screening tools” or “how to improve candidate ranking”, the direction is clear:

systems need to understand experience, not just detect keywords.

This is exactly the idea behind tools like Hero Parser.

Instead of scoring candidates based on keyword presence, it analyzes how those skills were actually used. It differentiates between short-term exposure and long-term, production-level experience.

This leads to rankings that reflect real capability — not just how well a CV is written.

The future of IT recruitment

The future of recruitment isn’t about avoiding AI.

It’s about adapting to it.

AI has standardized how candidates present themselves. That trend won’t reverse.

The only sustainable advantage now is better interpretation of candidate data.

Companies that move beyond keyword matching will hire faster and more accurately.

Those that don’t will keep asking the same question:

“Why do all candidates look perfect, but none stand out?”

Final thought

If you’re struggling with:

• too many strong-looking CVs
• low signal in candidate screening
• difficulty ranking applicants

you’re not dealing with an AI problem.

You’re dealing with an evaluation problem.

And that’s where the real opportunity is.

If you want to see what candidate ranking looks like when experience is actually taken into account, you can take a closer look at Hero Parser:

👉 https://heroparser.com

It’s built specifically for IT recruitment, where the difference between candidates isn’t in the keywords — but in how those skills were used in real projects.