You CAN automate a recruitment process until it stops recruiting..

Recently, I applied directly to a company for a CTO position. On paper, the match was unusually strong: software architecture, distributed systems, legacy modernization, technical leadership, team building and previous Acting CTO responsibility.

The first meaningful stage was an AI-led video interview. No conversation. No opportunity to clarify an answer, challenge an assumption, explore a trade-off or ask why a particular problem mattered to the business.

I did not pass.

Some time later, the same position appeared again, this time through a recruitment company. I was approached, went through the role carefully and invested considerably more than a day preparing for the discussion.

Then came the next step: exactly the same AI video interview.

I declined to repeat it.

Not because preparation is beneath a C-level candidate. Not because I am uncomfortable with technology - I have spent more than two decades building the stuff. I declined because repeating the same inadequate evaluation method and expecting it to reveal something different would not be recruitment. It would be ceremonial persistence.

I understand why companies need filtering. A popular position may attract hundreds of applicants, many of them completely unsuitable. Automation can remove obvious mismatches, verify basic information and help recruiters manage volume. That is a legitimate use of technology.

But a CTO is not recruited by verifying whether someone can deliver polished answers to a camera.

A substantial part of any C-level role - easily 40% in many situations - is not speaking. It is understanding. Reading hesitation before it becomes resistance. Hearing the uncertainty hidden behind an apparently clear statement. Recognising disagreement when everyone in the meeting is politely nodding. Distinguishing a technical problem from an organisational one and knowing when the person asking for a solution is not yet describing the real problem.

These are visual and auditory cues, but an asynchronous AI interview reduces them to a one-way performance. The candidate may be visible and audible; the organisation is neither.

A capable human interviewer can interrupt, probe, contradict and follow an unexpected line of reasoning. They can discover how a candidate thinks when the prepared answer ends. Just as importantly, the candidate can evaluate the company’s leadership, communication and decision-making. At C-level, that evaluation must work in both directions because the relationship being considered is not merely employment. It is shared responsibility for the company’s future.

Removing all meaningful human interaction from that process is therefore more than impersonal. It is itself a signal.

The company may believe it efficiently rejected an unsuitable candidate. It may equally have rejected exactly the person it needed, twice, without anyone ever speaking to him.

Automation can reduce the cost of recruitment. Used carelessly, it can also reduce the value of everyone who survives it.