The Boardroom Demo Problem: Building Recruitment AI That Actually Works in Production
There is a pattern Geert Tewissen has watched repeat itself across the staffing sector. A demo that impresses everyone in the room. An eighteen-year-old nephew with a laptop building something in two hours that would have cost a small fortune a decade ago. And then silence — because getting from a boardroom demo to a system that handles hundreds of thousands screenings a year, integrated with existing tools, running reliably at production level, is an entirely different problem. One that most players have not solved.
Taloo, co-founded by Tewissen and Laurijn Deschepper, is built around exactly that gap.
A Founder Who Has Lived Both Sides
Tewissen is not a pure tech founder who stumbled into recruitment. He started as a recruiter, ran his own agency, went through a bankruptcy that redirected his career toward service design and UX, and spent the following years building digital products and platforms for large clients at agencies including Nascom,Boondoggle and his own strategy studio, Superlab.
His minor in linguistics gave him a long-standing obsession with conversational interfaces. He spent years waiting for the technology to catch up. When it did, he was at Prato HR — a payroll and staffing software company with an entrepreneurial culture — and the conditions were right to build something new. That became Taloo.
His co-founder Laurijn Deschepper brings a complementary profile: an engineer with experience building products inside the Cronos Group. "He's a business-savvy engineer," says Tewissen. "I'm a tech-savvy business guy. We cover the same ground from different sides."
The Stagecoach Problem
Taloo's founding argument is best illustrated by one of their favourite slides: a patent drawing of a stagecoach with a steam engine bolted to the front.
"That's how new technology always gets adopted at first," Tewissen explains. "You slot it into the old paradigm and wonder why the efficiency gains don't materialise."
That, he argues, is where recruitment AI sits today — particularly in the staffing and interim sector. Companies are experimenting. Many have run pilots. Almost all have stalled at demo level, because the jump from "impressive prototype" to "runs reliably at scale across our existing systems" is enormous, and rarely visible from the outside.
The challenge is not just technical. It is also a perception problem. "A vibe coding workshop gives you a working CRUD application by the end of the day. That's genuinely impressive. But it's not what we're building. Explaining the difference is sometimes the hardest part of the sales conversation."
What Taloo Actually Does
Taloo is a digital colleague that gets work done inside the tools recruiters already use, from ATS and scheduling to payroll, and communicates naturally with teams and candidates through Teams, calendar, WhatsApp and voice.
The architecture is modular and agentic. What looks like a single "pre-screening agent" is, in practice, a coordinated system of several sub-agents. Tewissen organises the product roadmap around four areas:
Automated pipeline — pre-screening calls in 72 languages, CV and document collection, candidate qualification, 24/7 availability
Real-time assistance — prompting recruiters during intake and interview moments; time management nudges; ensuring the right questions get asked
Smart follow up — chasing candidates and hiring managers for responses, removing the manual back-and-forth cycle
Zero-effort adminstration — pre-checking timesheets and handling routine admin before it reaches a human desk
Human in the loop throughout. The recruiter controls the system; the AI removes the noise.
Integration runs via API to existing ATS systems. No replacement, no months-long IT project. Onboarding takes four weeks — a timeline Tewissen admits he questioned repeatedly until their first implementation proved it.
Taloo's first pilot with Itzu, a Flemish staffing company, is now live.
The practical flow: a vacancy is uploaded to the client's ATS as normal. Taloo's agent scrapes it, generates interview questions automatically, and activates when a candidate clicks "apply." The candidate receives a call. The screening happens. A score is produced. An appointment is booked. The recruiter finds a prepared agenda slot and a concise executive summary — without having touched the process.
"Even if we launched it in secret," Tewissen says, "the recruiter would notice on Monday morning. There would just be well-prepared appointments in the diary."
Where They Are on the Journey
The primary target market is staffing and interim agencies — high volume, margin pressure, and urgency. "The fat is off the soup," says Tewissen. "Efficiency gains have to come from somewhere, and these agencies feel that more acutely than anyone." Secondary markets are recruitment agencies and internal hiring teams, though he believes more evangelising is still needed there.
Beyond Belgium, conversations with imec are underway. The model is designed to scale, and Tewissen is clear that Flanders is the starting point, not the destination.
Why This Matters
The staffing sector does not need more AI demos. It needs AI that survives the move from presentation to production — domain-specific, cross-system, genuinely deployable by a team whose IT department is one person. That is exactly the problem Taloo is trying to solve.