Company Profile

Qonto, Europe's leading business-finance platform, serves 600,000+ SMEs and freelancers across 8 markets with around 2,300 employees. Its activation and onboarding team brings new business accounts live across the full setup cycle: KYC document collection, account configuration, compliance checks, and data entry. The team operates at very high volume, processing roughly 5,000 new business accounts every week across the eight markets.

Onboarding at Qonto isn't a multi-week consulting engagement. It's a regulated, document-heavy workflow where speed and accuracy both matter. Every account that takes too long to activate is an SMB that may have already moved on, and every error caught after activation is a compliance question that gets escalated.

As Qonto builds out its AI-native infrastructure across the platform, the onboarding team needed an agent purpose-built for the implementation layer: pulling document and KYC context, surfacing the next move on every account in flight, and running the manual setup work that used to compound into headcount.

Challenge

Onboarding specialists were losing the majority of their week to non-account work. The job description was "activate new business accounts." The actual day looked like coordination, chasing, and clerical work:

• Chasing customers for missing KYC documents and information, often through multiple back-and-forth cycles per account

• Manually reviewing and QA-ing submitted data for accuracy before activation, because customers regularly submitted incomplete or inconsistent information that wouldn't pass compliance

• Coordinating internally with compliance and risk teams on flagged accounts, including pulling supporting documentation and writing up context

• Logging status, customer notes, and case decisions across the CRM, often after the fact, often from memory

The math compounded against the team. Each specialist could only manage a limited book of concurrent onboardings because the per-account coordination work was real human time. As Qonto's customer acquisition grew, the only lever for keeping up was hiring. Volume growth meant a linear need for new specialists, and onboarding became one of the larger headcount lines in the operations org.

Onboarding also faced a quality ceiling. A specialist running too many accounts in parallel started missing things, escalating less rigorously, or letting accounts drift. Headcount was the safety valve. There wasn't another one.

What Tandem did

Tandem deployed without disrupting Qonto's existing CRM, compliance workflows, or specialist tooling. It sat on top, doing four things:

• Closed the customer follow-up loop. When a customer went quiet on a KYC request or submitted incomplete information, Tandem auto-drafted the follow-up message based on the specific item still missing. The specialist reviewed the draft, made any tone or context adjustments, and clicked send. What used to be ten minutes per account of writing personalized but essentially boilerplate messages now took thirty seconds of review.

• Validated submitted data automatically. When a customer uploaded their KYC documents, business registration, or compliance information, Tandem pulled the data, cross-checked it against required fields, and flagged only the exceptions that needed human review. Specialists stopped doing line-by-line QA on every account. Compliance-flagged data still routed to a specialist for judgment, but the long tail of routine validation work got absorbed by the agent.

• Scripted the compliance and risk handoffs. When an account got flagged for enhanced due diligence or a risk review, the handoff used to live in a Slack thread or an internal email. Now it's a tracked task with the right context attached, routed to the right reviewer, with a clear SLA and an automated reminder if it ages. Compliance reviews that used to take three to five days now run on schedule.

• Kept the CRM updated automatically. Tandem logged the customer interactions, document submissions, follow-up sends, and case decisions in the background. Specialists stopped spending the end of their day catching up on documentation. Leadership got accurate, real-time data on every account in flight.

The work that didn't move was the work that requires judgment: the borderline KYC calls, the compliance escalations, the customer relationships where context matters. Tandem absorbed everything around those moments, so the specialist had more time and attention left to make the calls that actually require a human.

Results

The 58% increase in accounts-per-specialist is the operational headline. The 43% reduction in admin time is what made it possible. With less time per account spent on coordination and logging, each specialist had real capacity for the next account, and the next, and the next.

The ROI math

• ~23 hires avoided × ~€80-100K fully-loaded cost ≈ €1.8-2.3M/year in payroll avoided

• Tandem cost for the ~90-person onboarding org ≈ €0.85M/year

• 2x to 2.7x return in year one, with net savings of ~€1M to €1.5M/year before counting the revenue impact of faster activation

This is the math that gets the renewal signed without a question. The headcount line for onboarding was projected to grow 34% over twelve months to keep up with customer volume. After Tandem, the same team absorbed the growth without a single new specialist hire. The savings are net of Tandem's cost, and the activation speed improvements (which compound into revenue recognition and SMB retention) haven't been priced in.

Why leadership cares

Onboarding had become one of the larger linear costs in Qonto's operations: a function whose headcount had to grow in lockstep with customer acquisition. Tandem changed the slope. The team scales now with the same number of people, even as volume grows. That's the difference between an operating expense that compounds with growth and one that decouples from it.

"We were hiring just to keep pace with growth. Tandem let the same team take on a third more volume. The headcount math changed completely."

— Martin Bouvier, Head of Customer Onboarding