Why AI Won't Capture the Best Talent 🪝
And What to Do Instead
When Zbynek Novotny and I brainstormed the title of our talk for Forbes, "The Best Talent Won’t Be Captured by AI," we worried people might label us as modern-day Luddites, the 19th-century workers who destroyed machines fearing job losses. But the reality is exactly the opposite: we love AI and use it daily. However, there are areas where AI simply can’t (just yet) match human judgment, empathy, and authenticity. Hiring is one such area.
The Inflationary Tsunami of Digital Communication
Do you remember November 30, 2022?
That date went down in history with the launch of ChatGPT, which forever changed digital communication. Since then, AI has literally moved the world.
One of the most concrete examples, and not just in recruiting, is the inflationary tsunami of cold emails. Data shows a 189% increase in AI-generated outbound. So, the quantity is growing incredibly fast. (Instantly.ai, 2024).
But at what cost? The open-rate of cold emails is dramatically falling – from 36% in 2023 to just 27.7% in 2024.
Why Candidates Recognize AI Emails and What Worries Them Most
We conducted our own LinkedIn survey to find out why this is happening. We focused on our outbound strategy, where we actively reach out to and seek talented senior candidates who are not currently looking for a job. The results were clear:
83% of candidates can identify an AI-written cold email.
What's even more important? Almost half, specifically 47%, wouldn't even respond to such an email!
91% of outreach emails are ignored because they feel generic.
While AI generates an immense quantity, the most important aspects – relevance and emotion – are lost.
We moved further into the process and asked what people fear most if AI were fully integrated into the recruitment process. People don't remember what you said – they remember how you made them feel. That's why the human touch in talent acquisition is crucial, even in the age of AI. Our data clearly shows that human contact is king:
86% of candidates are concerned about the loss of human connection and context.
69% see building trust and relationships with the team as the phase that AI replaces the worst.
From our data, where we asked over 1000 candidates about their biggest concerns regarding a fully AI-led initial interview, 49% feared the loss of human connection, and 37% feared missing nuance/context.
And regarding the very end of the hiring process – the crucial and emotional peak – the data is stark: 81% of candidates would not want an AI avatar to present their final job offer, finding it too impersonal.
The Paradox of AI-First Companies
We thought to ourselves: Okay, we have the data, we feel the emotions. But what about the best? Those who are furthest along in AI? If I want to be inspired and learn what to automate, I'll look at the most innovative companies, right? We went through the career pages of giants like OpenAI, Anthropic, Perplexity, Scale AI, and Inflection. Surely, their recruitment must be fully automated by now, with only robots working there?
And here came the paradox. What shocked us was that instead of seeing only AI algorithms and bots, we found that a big portion of open headcount in these AI-first companies consisted of the recruiting team! Most of their career pages were literally filled with roles like 'Recruiting Coordinator,' 'Executive Recruiter,' 'Recruiter, Public Sector.'
Today, we would assume that AI would take over the entire sourcing department, finding top talent in five minutes, and we wouldn't have to manually go through hundreds of profiles. The reality? As they told me at Ashby, the ATS vendor for OpenAI:
"Full automation of coordination runs into a huge amount of context that today's AI cannot yet grasp; it lacks memory, independent decision-making, and human judgment."
Even the most powerful AI needs "human glue"; without it, key details will be missed.
How to Stand Out in the Age of AI?
What to do about it? Stand out! That was true before AI, and it's true today. The unifying element of all the tactics mentioned here is the ability to stand out.
From my last few months of testing at least 10 recruiting AI tools and 10 agents, the things that worked best for me were, frankly, quite "cringe-worthy." For example, I had AI compose a customized song for an absolutely amazing product designer. I knew about one cold candidate designer from her website, that she loved cats and had a beautifully creative design portfolio in the world, and that was enough for my AI prompt. It was nerve-wracking, but it worked in 60% of cases! One of them is today the most talented designer at Attio.
If you need more "cringe" examples, do you know where AI really shined for me? In image testing! When I swapped my real photo for a funny AI avatar on LinkedIn, the response rate reached 60% of replies. I thought the avatar was boring, so I switched to a real photo. The response rate plummeted to 7%! So I decided to go back to the avatar, and it brought my response rate back up to 45% and is still growing. I take it a little personally, but one thing is clear: people react to the form before the content.
Transparency and the Human Touch
AI & Admit It: One of the most effective approaches is transparency. Signing off with 'Drafted by GPT-4, finished by Tereza' increased trust by +12%. Why does it work? People want to know what's truly behind the text. Transparency breaks the "black-box" feeling and gives candidates a sense that you're dealing with them fairly. External studies, like those from HBR, confirm this – openly labeling AI content maintains and often increases brand trust.
Handwritten Postcard: Zero spam rate, 100% open rate, and the candidate feels it's genuinely for them. This is how you can stand out and instantly gain attention and sympathy. As Jiri Semmler said, to whom our amazing Zbynek sent a postcard, even though he wasn't looking for a job, because of how it was done, his "first shot would go to Better Stack" if he were to change jobs.
Loom Video: Another tactic that really works for us is a 60-second Loom video for the candidate. Compared to a regular email, we get three times more replies – because the candidate sees a live face, hears a tone of voice, and immediately feels that it's meant "just for them".
Strong Brand as a Recruiting Weapon
An employer brand is not just a nice business card, but a direct recruiting weapon. Companies with an established brand attract up to +50% more qualified applicants, because candidates already know you, trust you, and want to be part of your mission.
How to achieve this? You can create podcasts (e.g., People of Productboard), share company stories on LinkedIn, and showcase the real environment in videos and blogs. The result? Higher trust = a better pipeline. The data proves it: 75% of people first Google a company before even responding to a recruiter's message , and 88% of candidates would rather accept less money if the company is known for its culture.
And at the end of the hiring process: 90-minute offer calls, where we share everything – revenue numbers, more about the team and how it works, about the product and a detailed future roadmap, our Series A fundraising deck. By doing this, you show interest and build trust.
AI will make good recruiters even better – and bad ones just faster.
Great post!
Regarding the avatar - isn't the change in response rate rather linked to your avatar looking like a Notion avatar and people being curious if they are getting an offer from Notion? You need to test it again with a different avatar 😂