When job applications become interchangeable – an interview with Florian Dyballa from Aivy

Job applications are becoming more professional. AI helps with CVs, cover letters and tailoring applications to specific roles. At the same time, some companies are receiving significantly more applications than they did just a few years ago. But what does this actually mean for recruitment? – I’m discussing this with Florian Dyballa from Aivy. Florian is a business psychologist, co-founder of Aivy and will be speaking at the SAATKORN Online Conference on 17 September on the topic of HR tech. Let’s get started:

SAATKORN: Florian, why don’t you start by briefly introducing yourself to SAATKORN readers?

I’m Florian Dyballa, a business psychologist and co-founder of Aivy. I built my first website at the age of 14 and founded my first company at 16. I actually came to business psychology through starting my own business. I wasn’t just interested in what was technically possible, but above all in what drives people, where their strengths lie and how best to recognise their potential. One idea continues to play an important role for me to this day: talent is evenly distributed, but opportunities, unfortunately, often aren’t. I’m a working-class child myself. The idea that I would one day go to university and set up a technology company was almost unimaginable in my family. There were no role models in my circle that I could have looked up to. That’s why I know from personal experience how easily potential goes undiscovered. A CV often shows, above all, what opportunities someone has had so far, but not necessarily what that person is capable of. That is a key reason why Aivy exists.

A CV shows opportunities – not necessarily potential

SAATKORN: Tell us a bit more about Aivy. How do you aim to make precisely this potential visible?

Aivy is a spin-off from Freie Universität Berlin. We help companies identify applicants’ strengths and cognitive potential at an early stage and in a structured way – that is, before the first interview. To do this, applicants complete short psychological tests on their smartphones. Each exercise takes just two to five minutes. Depending on the role, different tests are combined, so the entire assessment usually takes only about 10 to 20 minutes. We look at four areas: skills, personality, cultural fit and interests. This provides companies with a suitability score for the role in question and specific guidance for the rest of the selection process, such as suitable interview questions.

It is important to us that the applicants themselves also receive their personal strengths profile. Participants therefore receive feedback that they can put to good use even if they do not secure the specific role. More than 200 companies now use our software, including Lufthansa, Würth and Hermès. Around one million assessments have already been carried out via the platform. Our procedures are based on the quality requirements of DIN 33430, have been validated using established psychological tools and are regularly reviewed. Naturally, we also use them in our own recruitment processes.

More applications, less insight

SAATKORN: Your talk at the SAATKORN Online Conference is titled ‘When applications become interchangeable: How to make better recruitment decisions again’. What’s the idea behind it?

There’s one phrase I’m hearing in almost every conversation with HR managers at the moment: ‘We’re receiving more applications than ever before, but the selection process isn’t getting any easier.’ Behind this lie several developments that are currently converging: Easy Apply, remote working, the economic climate and, of course, generative AI.

However, the problem isn’t just the volume. At the same time, it’s becoming harder to tell from the application documents who is truly suited to a role. These days, almost anyone can have a professionally written CV generated in a matter of minutes, tailored precisely to the role. From the applicants’ perspective, this is understandable at first glance. For companies, however, this means that application documents are losing some of their previous significance. Nevertheless, many selection processes still rely heavily on CVs and cover letters. My session will therefore focus on where this approach is now reaching its limits, why some obvious responses do not solve the problem, and what a better pre-selection process might look like.

SAATKORN: How much has the situation actually changed – and why is it not enough simply to screen AI-optimised applications using AI again?

Essentially, there are four key points.

Firstly, I’ll show just how much the situation has already changed. Greenhouse has analysed more than 640 million applications from over 6,000 companies. The number of applications per recruiter per year has risen from 146 to 746 between 2022 and 2025. That is an increase of 412 per cent. At the same time, the average number of recruiters per company has fallen by 56 per cent. Furthermore, people who regularly use AI in their job search are more likely to be invited to interviews. This does not automatically mean that they are a better fit for the role in question. However, it does demonstrate the impact that professionally optimised documents can have on the screening process.

Secondly, there is the question of why AI screening alone is not a sufficient response to AI-optimised applications. AI can provide valuable support to recruiters. However, if it mainly analyses CVs and cover letters, it continues to evaluate information that is becoming increasingly easy to optimise. This speeds up the selection process, but does not automatically make it better. Furthermore, under the EU AI Act, AI systems that assess applicants or significantly influence selection decisions are classified as particularly sensitive.

Assessment needs to happen earlier

SAATKORN: If CVs and cover letters are losing their significance as the first selection indicator – what should happen earlier in the recruitment process instead?

Thirdly, I’ll show what changes when companies bring the diagnostic phase further forward. Typically, CV screening takes place at the start and an assessment towards the end of the process. This made sense as long as the number of applications could be effectively reduced through document screening. If this indicator becomes less reliable, it is worth reconsidering the order of events. In doing so, I also address the objection we hear most frequently: “Won’t too many applicants drop out then?” On our platform, we see completion rates of 85 to 95 per cent. A short assessment therefore need not be an insurmountable hurdle if it is designed fairly and the applicants themselves derive added value from it.

Fourthly, there are three interrelated points: clearly defining the requirements of a role, identifying relevant potential that isn’t apparent from the CV, and consistently basing the pre-selection process on a candidate’s suitability for the role. Technology can help with this. However, it does not relieve you of the task of determining in advance exactly what you are looking for.

SAATKORN: And what exactly can participants expect from your session on 17 September?

It’s important to me not to explain the topic purely in theoretical terms. That’s why I’ll be showing concrete examples from companies and letting them speak for themselves in short video statements. This will highlight just how different the initial situations were and what actually changed as a result of the shift in the selection process. However, I’ll also be discussing what diagnostic tools cannot achieve, where their limitations lie, and for which companies their use might not be worthwhile at all. So it’s not about presenting a single solution for every recruitment process.

The whole topic is also available as a white paper, complete with the studies and sources for further reading. And, of course, there’ll be plenty of opportunity for questions and comments in the live chat.

SAATKORN: Florian, thank you very much for the interview. I’m looking forward to your presentation at the SAATKORN Online Conference!

Further links:

Florian Dyballa on LinkedIn

Aivy website

SAATKORN Online Conference: register for free HERE!