Reinventing the First Interview
Fika Jobs, a Stockholm-based startup, raised 4 million dollars in pre-seed funding on June 23 to pursue an idea that is equal parts intriguing and unsettling, replacing the resume and the recruiter screen with a persistent video profile and an interview conducted by an AI agent. Candidates record a profile once and complete a roughly ten-minute interview powered by Google's Gemini, with questions personalized from their LinkedIn history. Rather than applying to jobs one at a time, they maintain a standing presence that employers can discover, inverting the usual flow of recruitment.
The founders, brothers Jakob and Alexander Dubois, frame the product as a fix for a familiar frustration. Recalling an earlier venture, Jakob Dubois said that when we were building the social app Gaff, we spent a lot of time recruiting and almost passed on a candidate because his resume did not really stand out. That anecdote captures the pitch, that resumes are a lossy format that hides communication skills and personality, and that a short structured video interview surfaces traits a document never could. Whether AI is the right interviewer for that job is the question the market will now test.
The Business Model Is the Real Bet
The pricing is where Fika Jobs makes its sharpest wager. The platform is free for job seekers, and employers pay 10 percent of first-year salary only upon a successful hire. That success-fee model aligns Fika with outcomes rather than activity, and it positions the company directly against traditional recruiting agencies that charge similar percentages for far more manual work. If an AI agent can do the early screening that agencies bill heavily for, the economics become compelling for employers tired of paying for effort rather than results.
We find the model more interesting than the technology, because it forces a clear value proposition. Fika only earns money when a placement works out, which means the company is betting its revenue on the quality of its AI screening, not just its ability to generate interview volume. That is a healthier incentive than the pay-per-post or pay-per-seat models common in recruiting software. It also raises the stakes for accuracy, since a platform that surfaces bad matches simply does not get paid, a discipline that should, in theory, keep the product honest.
Famous Backers and a Familiar Playbook
The round was led by Luminar Ventures with participation from Alliance VC, and it drew two names that resonate in the Nordic tech scene, Sebastian Knutsson and Riccardo Zacconi, co-founders of King, the studio behind Candy Crush. Their involvement is more than celebrity garnish. King built a business on engagement mechanics and consumer-grade product polish, and a video-first hiring product lives or dies on exactly those qualities. The candidate experience has to feel inviting rather than clinical, and that is the kind of design problem the King founders know intimately.
Early traction is modest but real, with more than 100 companies on the waitlist and over 50 having tested the platform ahead of a public launch expected in the fall. Fika plans to start in Sweden before expanding internationally, a sensible sequencing that lets it refine the product in a single labor market before confronting the regulatory and cultural variation of others. The 4 million dollar pre-seed is small by current AI standards, which is appropriate for a company that still has to prove its core claim, that candidates and employers will trust an agent to run the first interview.
The Uncomfortable Questions About AI Interviewers
We cannot cover this without naming the obvious concerns. AI-conducted interviews raise hard questions about bias, fairness, and transparency, and the regulatory environment is tightening around exactly this use case. Several jurisdictions now require disclosure and bias auditing when automated systems are used in hiring decisions, and an agent that generates personalized questions and evaluates responses sits squarely in scope. Fika's reliance on a general purpose model like Gemini means its fairness properties depend partly on a system it does not control, which is a real governance wrinkle.
There is also the candidate trust problem. Many job seekers already resent automated screening that rejects them without a human ever looking at their materials, and putting an AI in the interviewer's chair could deepen that resentment if it feels like talking to a wall. Fika's counterargument is that its approach is more humane than the resume keyword filters it replaces, giving candidates a chance to show personality rather than be screened out by an applicant tracking system. That is a defensible claim, but it is the company's central burden of proof, and the market will be unforgiving if the experience feels dehumanizing.
A Small Round With a Big Thesis
Fika Jobs is a minor funding event in dollar terms, but it is a clean illustration of where agentic AI is heading in the enterprise, into the operational workflows that companies have always found tedious and expensive. Recruiting is a perfect candidate, high volume, repetitive in its early stages, and costly when done by agencies. If agents can handle the first pass reliably and fairly, the savings are substantial, which is why we expect to see many more startups aiming agents at specific business processes rather than at general assistance.
The broader lesson for technology leaders is that the agentic wave will arrive function by function, often through small specialized vendors rather than sweeping platform launches. Hiring, procurement, support, and finance operations are all being targeted by companies with theses much like Fika's. The winners will be the ones that pair credible AI with a business model that aligns incentives and a product experience that people actually tolerate. Fika has a sharp model and famous backers, and now it has to prove that an agent can earn the trust a first interview requires.
Why HR Is an Early Agentic Battleground
Hiring is emerging as one of the most contested frontiers for agentic AI precisely because the stakes are human and the inefficiencies are glaring. Recruiting teams spend enormous effort on early-stage screening that is repetitive yet consequential, and the economics of automating it are obvious. That is why we see a wave of startups, Fika among them, aiming agents at specific corners of the talent lifecycle, from sourcing and screening to scheduling and reference checks. The category is attracting capital because the cost base it targets is large and the manual work is genuinely tedious.
The flip side is that hiring is also where AI's social and legal risks are most acute, which means this battleground will be shaped as much by regulators and candidates as by technology. Employers experimenting with agent-led interviews should insist on transparency, bias auditing, and a clear human appeal path, not only to comply with tightening rules but to preserve trust with the people they hope to hire. The vendors that win durable enterprise business will be those that treat fairness and explainability as features, not afterthoughts. Fika's success-fee model gives it some incentive alignment, but the governance bar in HR is unforgiving.
