Xavier Dekeyster

Xavier Dekeyster's public voice centers on one conviction, tested repeatedly in practice: AI-driven planning proves its worth precisely where reality is messiest. His clearest evidence comes from a real-world trial in which a client's planners built schedules manually for 125 days, after which FieldGenius AI was asked to replan the same period blind, without knowledge of the choices made. The result, as he describes it, was that "op rustige dagen was de AI-planning minstens even sterk als de manuele planning," but "op drukke dagen met veel wijzigingen, afhankelijkheden en uitzonderingen maakte Fieldgenius duidelijk het verschil" . From this he draws a general principle he returns to often: "hoe complexer de puzzel werd, hoe sterker de AI presteerde" . This isn't a one-off observation but the foundation of how he positions the product itself: "FieldGenius is niet gebouwd voor eenvoudige planningen. Het is gebouwd voor organisaties waar de realiteit voortdurend verandert" .

That framing of complexity as the real test of technology recurs across his writing on FieldGenius. He's careful to define what the product is and isn't, resisting easy categorization: "Is FieldGenius een routeplanner? Ja. Is het een CRM? Ook. Kan het werkorders beheren? Zeker. Maar tegelijk is het geen van die drie" . Instead he describes it as "een AI-gedreven intelligentielaag die bovenop je bestaande software werkt met of zonder UX/UI," pulling together customers, work orders, technicians, skills, materials, SLAs and availability to compute an optimal schedule in seconds . Crucially, he treats a finished schedule as already obsolete: "Een planning is al achterhaald op het moment dat ze klaar is," which is why the system needs to keep adjusting for traffic, weather and other live variables rather than producing a static plan . His interest in weather as a planning parameter — accounting for rain, wind, heatwaves and regional warnings — reflects the same underlying belief that a good schedule must be dynamic and context-aware, not fixed at the moment of creation .

Dekeyster is equally vocal about the practical friction of building with AI himself, and here his tone becomes notably more measured than promotional. After experimenting extensively with tools like Claude Code, GitHub Copilot, Cursor, Windsurf, Lovable and Bolt, he concludes "de resultaten zijn indrukwekkend," but immediately qualifies it: "de eerlijkheid gebiedt mij ook te zeggen dat het niet altijd zo vanzelfsprekend is als sommige demo's doen uitschijnen" . He points to specific failure modes — deployments that keep failing, integrations that almost work, recurring configuration problems — and admits "meer dan eens heb ik zelf rechtstreeks in de code moeten ingrijpen om verder te geraken" . He attributes his sensitivity to these limitations to the fact that DX-Solutions builds software and AI solutions itself, which lets the team spot where the boundaries currently lie faster than an outside user might . This combination of enthusiasm and candor about rough edges is a consistent pattern: he promotes AI's capabilities while explicitly refusing to oversell its maturity.

On enterprise architecture, he articulates a clear, considered position in the long-running suite-versus-best-of-breed debate. Rather than picking a side, he argues that "de meeste bedrijven uiteindelijk ergens tussen beide uitersten uitkomen" and that "niet de software zelf bepaalt het succes, maar de architectuur erachter" . He frames this as newly urgent because AI increasingly needs access to data, processes and knowledge across systems, which raises the stakes of how that architecture is designed regardless of vendor choice . This systems-level thinking connects to his broader professional identity: DX-Solutions' positioning around "meer vrijheid uit hun IT-landschap" and its work in middleware, integrations and custom development suggests architecture, not any single tool, is his real preoccupation .

Validation matters to him, and he's explicit about the bar it represents. Describing VLAIO's approval of the FieldGenius development project, he stresses "een VLAIO ontwikkelingsproject krijg je niet op basis van een goed idee alleen," listing the requirements as "een aantoonbare kennisuitdaging, een sterke technische onderbouwing, een realistisch valorisatietraject en een duidelijke economische impact" . He traces FieldGenius's origin directly to repeated client pain points DX-Solutions encountered over years of software and integration work, where planning — availability, skills, vehicles, materials, travel time, urgent interventions, geographic spread and customer preferences — kept surfacing as one of the most complex operational processes organizations face . In other words, the product is presented not as a speculative bet but as a direct response to a problem he says he saw "telkens" .

Beyond the product narrative, Dekeyster occasionally steps into broader commentary, and when he does it's grounded in his own experience as a long-time entrepreneur. Addressing a political discussion about wealth taxation, he lays out, in plain terms, how paper valuation of a company doesn't translate into personal liquidity: "Ben je dan plots miljonair? Nee. Dat miljoen staat niet op je privérekening. Het zit in je onderneming," pointing to buildings, machines, people, inventory and future growth as where that value actually resides . He frames this as something he doesn't usually do — "ik probeer me normaal niet te veel uit te spreken over politiek" — but felt compelled to address "als ondernemer van meer dan 25 jaar" . The concrete takeaway across all his material is consistent: he trusts systems and architecture over hype, values evidence from direct trials over demos, and speaks from an operator's vantage point whether the subject is AI planning, tooling limitations, IT strategy, or economic policy affecting business owners.