Automating first-stage screening without losing the human touch

What you're actually automating

Most resistance to automated screening comes from a misunderstanding: the assumption that an algorithm decides people's fates. That is not what happens here.

An automated screening conversation asks questions. The same questions a recruiter would otherwise ask in a ten-minute phone call – about availability, willingness to travel, relevant experience, expectations about the role. The conversation happens via WhatsApp, at a time that suits the candidate. The responses come in structured and comparable.

You then receive a profile: the conversation itself, a CV if provided, a score against the job criteria you specified, and suggested questions for the follow-up interview. That follow-up interview is yours to conduct. In person. With someone you already know more about than you would after the average first impression.

The question is not: do I automate, or not? The question is: which part of the contact deserves my time – and which part is really just information logistics?

The trade-off that matters

Automation saves time. That is the most familiar reason to consider it. There is also a less-discussed argument.

An automated conversation approaches every candidate consistently. Everyone receives the same questions, in the same order, without the recruiter's mood at ten to five affecting anything. That is not just efficient – it is also fairer.

At the same time, an automated conversation has limits. It does not pick up on tone, on hesitant enthusiasm, on the moment someone says "I've done something like that" in a way that invites a follow-up question. That is what you do in the follow-up interview.

The risk runs in both directions. Bring in personal contact too early and you spend time on candidates who drop out at the first basic questions. Leave it too late and you lose candidates who do not feel seen. The balance: gather information quickly, follow up quickly, and schedule the personal conversation quickly for those who clear the first filter.

What this looks like in practice

Suppose – hypothetically – that you are looking for a service engineer for installations in the manufacturing industry. The role requires a driving licence, a technical background, and a willingness to work across different regions. An automated conversation puts those questions to the candidate, follows up based on their responses, and gives you a clear picture of who genuinely fits the bill. Anyone who does not meet the threshold criteria drops out – without you having spent any time on them. Anyone who does clear the first filter, you invite to a conversation where you already know what you are talking about.

The same principle applies to a production planner with scheduling experience, or a technical project manager where attitude towards collaboration carries at least as much weight as the CV.

Step by step

1. Define the job criteria clearly – not as a wish list, but as genuine threshold criteria. What makes a candidate good enough to progress to the follow-up interview?
2. Pass those criteria through to the screening conversation. The AI recruiter follows up from that context, outside office hours too.
3. Wait for the profile – conversation, score, suggested questions – in the Cockpit.
4. Make your own assessment on whether the profile fits. The score is a tool, not a verdict on suitability.
5. Schedule the follow-up interview for candidates you find interesting. This conversation is yours: informal, personal, based on what you already know.

---