Fixing recruitment with Onceover AI: A chat with Sem Gabelko
Today in this interview, we have the pleasure of speaking with Sem Gabelko, the founder and CEO of Onceover AI. Sem is the driving force behind this innovative platform that is revolutionising the recruitment process through advanced AI technologies. Onceover AI aims to streamline and enhance the hiring process by using cutting-edge tools to match the right candidates with the right opportunities. We are excited to delve into Sem’s journey and learn more about the impact of his work.
TFS: Welcome, Mr. Sem Gabelko! Thank you so much for joining us today. We are really excited to hear more about your journey—what led you to start Onceover AI and how your platform is making waves in the recruitment industry. We hope to learn about your innovative use of AI in this important field and hear more about your story.
Sem Gabelko: Thank you for having me. It is a pleasure to be here and share my story. Building Onceover AI has definitely been a challenging journey, but it’s also been incredibly rewarding. I’m truly passionate about transforming the recruitment industry with AI. I’m excited to discuss our journey so far and where we’re headed next.
TFS: First things first, what inspired you to start your own business? How did the idea come to you?
Sem Gabelko: I have wanted to build my own business since I was about 17, if I recall correctly. The idea of personal growth, financial freedom, and the chance to create my own story and path was incredibly appealing. However, I was always hesitant to take the leap, feeling that I needed more experience to be better prepared for the journey. But as it turns out, you are never fully prepared to start your first company.
The idea for Onceover AI came to me during my first job as an ERP implementation consultant. I was tasked with optimising and digitalising business processes through ERP software. I found it frustrating because the implementation process was often very long and costly and did not always deliver the desired results. It might solve some issues, but often creates new ones as well. Because of the AI boom due to advancements in language modeling, I decided to look for ways I could apply the new technology to add true automation to business processes. I decided to focus on recruitment because it is a very language intensive and non-deterministic process. A perfect application for LLMs.
TFS: Now that we have discussed how Onceover AI started, I am interested to know what drove you to focus on recruitment with your platform. How do you see AI shaping the future of recruitment in the next five years?
Sem Gabelko: Recruitment in many companies is a subjective process. The screening phase tends to be quite rigid, which often means the best candidates might not get through, and only a few are selected for interviews. It usually comes down to hiring the person who seems the most appealing. Many companies lack a formal, tailored recruitment process, which leads to inefficiencies and poor fits.
Recruitment should be seen as the most fundamental business process and also the most important one because all organisations in the end are a combination of people. I believe that placing the right people in the right position at the right time is what separates successful companies from the rest. That is what we aim to standardise and democratise with AI so that every company can find the right people for them.
Looking ahead, the next five years will be about exploring and adopting new methods and technologies. AI will significantly improve speed and efficiency, while reducing cost and increasing the quality and accuracy of recruitment. However, AI will not take over everything. It will handle the tedious tasks, while recruiters and managers will still conduct the final interviews and evaluations and make the ultimate decisions. This way, they will have access to a much larger talent pool, reduce bias, and find the right candidates in the blink of an eye. But the final choice will always be theirs.
TFS: That sounds like a promising vision. Could you tell us more about the specific AI technologies and methods your platform uses? How do these set Onceover AI apart from the traditional recruitment tools out there?
Sem Gabelko: Our platform uses advanced large language models (LLMs) and combines them with traditional natural language processing (NLP) techniques. In the first phase, our system profiles and ranks resumes, which helps us sort through applicants more effectively. In the second phase, we conduct adaptable AI interviews with follow-up questions to get deeper insights before finalising the overall ranking. This approach allows us to match candidates more closely with both the job requirements and the company culture.
Traditional recruitment tools, on the other hand, lack this level of AI integration. They offer extensive digitalisation features to create a central repository of information, but most of the processes remain manual. This means recruiters still have to sift through resumes, schedule and conduct initial interviews, of which most end up in rejection. Also, conventional tools use Boolean searches based on keywords, which can lead to missing out on excellent candidates who do not use the exact terms in their resumes. On the other hand, you might end up with a pool of candidates who only appear to be a good fit because they match the keywords but are not necessarily right for the role. Our platform addresses these issues by improving accuracy, ensuring that the candidates selected are well-suited to the positions and the company’s environment.
TFS: That is impressive. How does Onceover AI adapt to the different recruitment needs of various companies? Could you share an example where you had to manage a particularly challenging customisation and how your platform managed it?
Sem Gabelko: Every company has its own unique needs when it comes to finding the right talent—not just for the job itself but also for fitting into the company’s culture and environment. To manage this complexity, we follow a few key steps. First, during the onboarding phase, we require new customers to complete a questionnaire that helps our system understand the company’s culture. Then, when a specific job is being created, our AI assists in crafting the job description with adaptable questions that result in a scorecard for evaluating and ranking candidates. You can specify the qualities that are essential for the job and how much weight each one carries. Finally, there’s an optional phase designed to enhance our model’s accuracy, where we gather more detailed information about various departments, management styles, and even past team-building activities.
TFS: That makes a lot of sense. As your platform becomes more advanced, how do you keep the user interface of Onceover AI simple and user-friendly? What steps do you take to ensure that it remains intuitive even as new features are added?
Sem Gabelko: We start by following good practices and design patterns that users are already familiar with. That is a solid foundation. However, the only way to truly ensure that is by regularly taking feedback sessions with your users. Whenever we add new features or make changes, I make it a point to mention them and ask users for their thoughts. Continuous feedback is often overlooked but is essential, especially as you grow. Making customers feel that their needs and opinions are valued builds trust, and in the B2B market, trust is everything.
TFS: Given the sensitive nature of recruitment data, what measures does Onceover AI take to ensure the privacy and security of both applicants’ and companies’ information?
Sem Gabelko: You are right; recruitment data is sensitive. That is why we follow best practices in software development, data management, and cloud architectures to maintain security. We also make sure we comply with GDPR guidelines, ensuring privacy by using the data we gather solely for the specific purpose of ranking candidates against the jobs.
TFS: How does Onceover AI manage to scale for companies of all sizes, whether they are startups or large enterprises? Could you share some of the challenges you have encountered while scaling your solution?
Sem Gabelko: Great question, but we have not reached that stage yet. At the moment, we have users who are small companies and startups. We always consider scalability when adding new features, as we aim to be forward-thinking. Of course, some issues will only become apparent once we are fully in production. We do our best to anticipate these challenges to avoid being caught off guard.
TFS: It sounds like you are well-prepared for future challenges. How does your platform make sure that AI recommendations and decisions are transparent? And how do you tackle any concerns about potential biases in the AI?
Sem Gabelko: Explainability is a big focus for us. We generate detailed reports for each job position and candidate, explaining the rankings and why a candidate might or might not be a good fit. In a future release we are planning to include reports that will be accessible to the candidates, including tips and feedback. While AI bias is a concern, it is more predictable and manageable compared to the bias that can occur when humans quickly scan resumes. Reduced bias compared to human evaluators is one of the advantages AI offers.
TFS: What key performance indicators do you use to see how well Onceover AI is working for your clients? How do you measure improvements in the recruitment process?
Sem Gabelko: We focus on the main metrics like time to hire, cost per hire, candidate pipeline throughput, candidate drop-off rate, offer acceptance rate, interview to hire and candidate experience. For the longer term we aim to track quality of hire and the retention rate as well.
TFS: Can you share a success story or feedback from a client who has significantly benefited from using Onceover AI? What was their experience with the platform?
Sem Gabelko: Of course. The main advantage we provide right now is saving both time and money. Our beta users, who often have limited recruitment resources, are experiencing these benefits firsthand. Take our first user, for instance—a young recruitment agency specialising in software engineering talent. They receive numerous requests but have limited resources for finding the right candidates. With our platform, they can effectively meet client needs without the financial strain of scaling their operations.
TFS: What exciting new features or improvements are you working on for Onceover AI? How do you plan to keep pushing the envelope in the recruitment space?
Sem Gabelko: Our upcoming milestone is a broad release planned for September. Right now, we are focused on refining the platform based on user feedback. Two key features we are working on include a more detailed onboarding process to better capture company culture and an explainability report for candidates to enhance their experience. By staying current with recruitment best practices and investing heavily in the R&D of our AI systems, we aim to remain at the cutting edge of innovation.
TFS: Now, let us shift focus a bit. What does a typical day in your life look like as an entrepreneur?
Sem Gabelko: As we are still in the early stages of the startup, preparing for a widespread release in September, my typical day involves a lot of hands-on work. I spend time researching, building and testing the product, and communicating with our beta users. There is also a lot of work involved in creating decks and presentations and working on building the brand. It can be quite detailed and demanding, but it is all part of laying the groundwork for our success.
TFS: How do you stay motivated and keep pushing forward, especially when things get challenging?
Sem Gabelko: I have developed a mindset where giving up is simply not an option and a belief that I have no other choice. I have always felt the drive to do more and achieve more, so it is less about staying motivated and more about discipline. I see each challenge as a necessary step towards reaching my goals and dreams.
TFS: That sounds like a great approach. As we wrap up, I would love to hear your advice for those who are thinking about starting their own business. What would you say to aspiring entrepreneurs?
Sem Gabelko: First, try not to overthink things or prepare excessively because you will never be ready. Just dive in and learn as you go. Second, avoid the trap of doing everything on your own. It is possible, but it is highly inefficient and can become very demotivating. Find people who share your ambition and complement your skills. Lastly, ensure you either really love problems and challenges, or you are deeply unsatisfied with your life and want to change it. Do not embark on this journey because of the hype or because entrepreneurs are regarded as legends or mythical creatures—we are not.
TFS: Excellent. What has been the most rewarding part of your journey as an entrepreneur? What makes you feel most proud?
Sem Gabelko: The most rewarding part for me is making my own decisions and charting my own course. At this stage of my journey, that is the only rewarding aspect.
TFS: This has been an incredibly insightful interview, Sem. It has been great learning about your journey and Onceover AI. Best of luck with your future efforts and thank you for joining us today.
Sem Gabelko: Thank you. It has been great talking about Onceover AI and sharing our vision. I appreciate the opportunity and your kind words.