Every day, 2.7 billion people wake up and go to work in roles that can't be done from a laptop. They drive trucks, stock shelves, cook food, clean offices, and deliver packages.

Without them, society stops.

Yet when these people look for jobs, they encounter a painful process: applications that require 20+ clicks and aren’t optimised for mobile, weeks of silence followed by generic rejections, and interview processes designed for people with flexible schedules and reliable internet.

The disconnect runs deeper than user experience. The people building hiring software live in one world. The people using it live in another entirely.

Eric André lived in both. Now, Cherry is backing him and the Kiku team as they rebuild hiring infrastructure for how frontline work actually operates.

The Frontline Hiring Reality

Before starting Kiku, Eric was part of the early team at Voi, joining when the company was just 20 people and helping scale it into the European category leader. He hired hundreds of frontline workers in the process across multiple countries.

He discovered that frontline jobs account for over 80% of the global workforce yet recruitment tools have largely overlooked this segment.

These roles face turnover rates of 60-80% annually, yet it still takes over 40 days on average to fill a role. Recruiters spend their time on repetitive tasks like screening and scheduling, while up to 80% of candidates report being ghosted in the process.

Most hiring tools are built for corporate environments, not for today's mobile-first, often deskless workforce. In a fast changing market where bots flood application systems, and companies compete globally for frontline talent speaking diverse languages, traditional hiring platforms fall short.

Traditional ATS (applicant tracking systems) assume low-volume, high-touch hiring where human judgment scales linearly. But frontline recruitment requires batch processing at machine scale while maintaining individual assessment quality. The existing infrastructure couldn't handle this fundamental contradiction.

Eric realised the entire category was addressing the wrong technical challenge.

Sara and the Architecture of Scale

Eric, joined by CTO Rasmus Andersen (a YC alum and serial founder with two exits, previously CTO at Trendsales, now part of Vinted) rebuilt the entire technical stack around volume hiring principles.

Kiku's AI agent, "Sara," operates through conversational voice and text interfaces designed for mobile-first candidates. Instead of lengthy forms or scheduled calls, candidates can apply and interview instantly - anytime, anywhere - through natural conversations that adapt to their pace and preferred communication style.

Early results across the Nordics, UK, and US are validating the approach -

• 93% candidate satisfaction rate.
• 75% faster time-to-hire.
• 15+ hours saved per role.

But numbers only tell half of the story. The challenge is rapidly identifying qualified candidates from massive applicant pools while maintaining consistent quality standards across geographies, languages, and role types.

This requires fundamentally different algorithmic design. Instead of deep personalisation models, you need robust classification systems that work across demographic variations. Instead of complex scoring algorithms, you need binary qualification gates that scale infinitely.

Sara handles code-switching, regional dialects, and the kind of authentic communication that traditional screening tools penalise. The models operate as a true multi-tenant system where each client's hiring criteria become training data that improves the overall platform.

The infrastructure maintains state across multiple interaction types. A candidate can start an application on mobile, continue via SMS, complete screening through voice, and finish onboarding in-person, all while maintaining context and data continuity.

Network Effects in Hiring Infrastructure

As Kiku processes more interviews, their models learn what questions and answers predict successful hires for specific industries and companies. This creates increasingly precise matching algorithms that improve placement quality while reducing time-to-hire, building competitive advantages that scale with data.

More powerful is the demand-side network effect. When candidates don't get hired by one client, Kiku can immediately match them to similar roles at other clients. Rejection at Company A becomes qualification for Company B, dramatically improving candidate experience while increasing platform utilisation.

As Kiku continues to scale, they're building a dataset of conversation patterns and hiring outcomes that becomes increasingly difficult to replicate. They're already seeing this advantage emerge with their current weekly interviews across multiple languages, industries, and role types - each interaction teaching their models what successful placements look like in real hiring scenarios.

Every successful placement creates a feedback loop that enhances matching quality across the entire network. The flywheel accelerates as more companies join the platform.

What excites us most is Kiku's potential to become hiring infrastructure rather than just hiring software.

As they expand beyond screening into sourcing, they can leverage placement data to predict which passive candidates might be interested in specific roles. As they integrate deeper with payroll and onboarding systems, they can reduce time-to-productivity for placed candidates.

Each expansion creates more touchpoints with the hiring workflow, more data about candidate preferences, and more switching costs for clients who integrate their processes around Kiku's infrastructure.

Why We're Backing Them

For Cherry, Kiku represents the convergence of several investment themes we've been tracking.

The application of AI to genuine business problems rather than technological novelties. Kiku's AI solves real workflow inefficiencies with measurable economic impact, not abstract capability demonstrations.

The opportunity to build infrastructure in underserved markets. Frontline hiring represents a massive addressable market that's been systematically ignored by venture-funded innovation.

The potential for European companies to compete globally in AI applications by focusing on specific use cases rather than general capabilities. Kiku's domain expertise and technical focus create advantages that generalist competitors can't easily replicate.

Eric and Rasmus understand both the technical complexity and the business dynamics of their market. They've built systems that work in the real world under actual pressure, not just in demonstrations under controlled conditions.

We're proud to back Kiku as they build the infrastructure that makes hiring work for the people who keep the world running.