For years, the corporate recruitment pipeline has behaved like a clogged drain. Job seekers spend countless hours tailoring resumes, polishing cover letters, and conforming to arbitrary templates, only to watch their applications disappear into the digital void. We call it the black box of Applicant Tracking Systems (ATS). But recently, this situation has gotten significantly worse.
The culprit is generative AI. With the floodgates wide open, candidates are now using automated tools to blast out thousands of hyper-optimized, machine-written resumes to every open role. In response, talent acquisition departments have scaled up their own automated screening tools to defend against the deluge. It is a machine-versus-machine war of attrition where the human candidate is entirely erased. When every resume looks perfect because a large language model wrote it, none of them actually tell you who the person is.
From a system migration perspective, building heavier filtering pipelines and stacking more screening software on top of old architectures feels like a band-aid. The core format is broken. The traditional resume is a legacy data structure—highly structured on paper, yet completely devoid of signal regarding a candidate’s grit, communication style, or adaptability. The more automated we make the old system, the faster we generate noise. We need to migrate to a new data schema that cannot be easily spoofed by a prompt. If you've spent any time managing engineering migrations, you know that trying to parse poorly structured, inflated data feeds always leads to garbage results down the line. It's the classic 'garbage in, garbage out' trap, just with human careers at stake.
Enter the Video-First Sieve: Inside Fika Jobs' Agent Auditions
Stockholm-based startup Fika Jobs is attempting to disrupt this paradigm by replacing text documents with interactive, video-first profiles. Founded by brothers Jakob Dubois and Alexander Dubois, the company recently announced a $4 million pre-seed round to launch a platform that looks like a hybrid between LinkedIn and TikTok.
The founders' inspiration came from a practical bottleneck when they were building their previous social app, Gaff. Jakob Dubois, Fika's CEO, admitted to TechCrunch that they almost passed on an applicant because his resume was unremarkable. They decided to speak with him anyway, and within minutes, they realized he had the exact grit and drive they needed. That realization highlighted a fundamental truth: paper resumes hide talent.
Fika’s solution is to replace the resume submission with an automated, Gemini-powered video interview. The candidate connects their LinkedIn profile, and the system generates custom interview questions based on their background. The candidate then participates in a ten-minute conversation with an AI agent. The platform cuts this video into brief, searchable clips and stores it as a live candidate profile.
Unlike enterprise-facing recruitment systems that screen applicants for specific employers on a per-job basis, Fika maintains a candidate-owned pool of pre-interviewed talent. Employers can browse this pool and filter candidates based on cultural alignment, remote work stipends, holiday expectations, and salary demands.
This candidate-owned pool model contrasts sharply with frontline workforce automation tools. For instance, Sergi Bastardas co-founded Orbio AI to solve frontline worker hiring and onboarding logistics under a multi-million dollar Series A banner. While frontline systems focus on high-volume logistics and instant scheduling, Fika targeting white-collar knowledge workers requires a deeper evaluation of communication, soft skills, and culture. Storing candidate interviews in a shared, discoverable repository rather than a siloed silo helps prevent job-seekers from repeating interviews for every application. It’s an interesting architectural choice: rather than running a fresh pipeline execution for every job poster, you cache the candidate's interview state and serve it to multiple consumers.
The Pre-Seed Funding & The Economics of Modern Recruiting
Building and scaling a video-first platform that processes hours of high-definition video through LLMs is computationally expensive. To finance the development, Fika raised a $4 million pre-seed round led by Luminar Ventures, with Alliance VC participating. The round also drew support from King co-founders Sebastian Knutsson and Riccardo Zacconi, the creators of Candy Crush.
The investors are likely betting on Fika's disruptive pricing model. Rather than charging high software-as-a-service subscription fees or the usual 20% to 30% placement fees demanded by traditional recruiting firms and headhunters, Fika charges employers a flat 10% fee on the successful hire's first-year salary. The service is free for candidates.
This transaction fee model shifts the financial risk away from the employer until value is delivered. For companies trying to manage cloud costs and tighten their budgets under current macroeconomic pressures, a contingency fee that cuts standard recruiter margins in half represents a compelling migration path. It aligns recruitment costs directly with successful outcomes. It's pragmatic business logic: if the match doesn't work, the treasury doesn't leak. Traditional recruiters who rely on expensive retains are going to struggle to compete with this automated cost structure. This shift reflects a broader trend of corporate resource reallocation; while some organizations focus on automating the pipeline for new talent, others are redirecting capital to prepare their current workforce for the AI transition, notably highlighted by Autodesk's AI upskilling initiative.
The Broader AI Interview and Video Screening Ecosystem
Fika Jobs is not operating in a vacuum. The recruitment tech ecosystem is crowded. Prominent platforms like Interviewer.AI and Hirevue have established strong footprints in the enterprise market by utilizing conversational AI and structured video screening. These platforms promise to reduce manual resume-parsing efforts by 60% to 80%, allowing HR teams to skip the initial phone screen entirely.
However, their strategies diverge in critical ways. While Fika targets early-career discovery through candidate-owned portfolios, legacy players focus on heavily quantitative features. Hirevue, for example, combines video screening with Virtual Job Tryouts, skill validation assessments, and AI-led mock interview practice to give candidates a chance to test their skills before facing hiring managers.
Beyond that, enterprise platforms operate as private vendor tools. A candidate applying to three different companies using Hirevue must complete three separate interview pipelines. Fika’s centralized candidate pool model seeks to eliminate this redundancy. By letting candidates own their pre-recorded interviews, Fika aims to reduce the fatigue associated with repetitive initial screenings, saving time for both candidates and employers. In simple system design terms, we are talking about moving from an expensive point-to-point architecture to a centralized event bus model where the candidate's state is broadcast once and ingested many times. While these platforms attempt to optimize volume screening, securing executive and highly technical roles remains a high-touch strategic endeavor, as seen in the hiring strategies of elite firms preparing for public markets, such as OpenAI's high-profile hires.
The Legal, Bias, and Ethical Minefields of Video-First AI Sifting
Despite the efficiency gains, migrating recruitment to a video-first model introduces substantial legal and ethical challenges. Resumes may be dull, but they provide a buffer. A text document obscures a candidate’s race, gender, accent, age, and physical appearance. When an employer’s first point of contact is a video clip, bias can influence decisions long before qualifications are reviewed.
This risk is why many modern enterprises have adopted blind resume screening, stripping names and demographic markers from applications to ensure objective evaluation. A video-first platform does the opposite, bringing demographic details to the forefront of the filtering process. This exposure creates a major compliance hazard for international companies subject to strict EEOC regulations in the US or anti-discrimination laws in Europe.
To address these concerns, major players in the AI recruiting space stress the importance of explainable AI. Systems like Interviewer.AI and Hirevue attempt to counter bias by auditability and providing objective breakdowns of scoring metrics. These systems focus on semantic analysis and pre-defined skill markers rather than subjective visual indicators. If Fika Jobs hopes to expand from its initial launch in Sweden into broader, highly regulated markets, the Dubois brothers will need to show that their AI agents evaluate candidates fairly and that their curated video profiles do not inadvertently systematize hiring bias. They must build robust guardrails, not just slick user interfaces. In my experience with legacy migration, the prettiest frontend is worthless if your backend database is leaking toxic compliance failures.