Christophe Ricevuto

The through-line

Across nearly every post, Christophe Ricevuto returns to one preoccupation: dependency, and how to eliminate it, whether it is dependency on a single employee, on an integrator, or on stale, scattered data. Early posts frame this around Odoo as a way to centralize information so a business does not collapse when one person leaves or one document goes missing [4][12][15]. As the month progresses, the argument sharpens into an explicit critique of his own industry, arguing that integrators profit from keeping clients dependent and that the right approach is to train clients into autonomy [2][8]. By mid-to-late August the same logic gets applied to a new frontier, AI plugged directly into Odoo, where he insists the real bottleneck is not the technology but the cleanliness of the underlying data [6][13]. The position has not really changed, but the target has widened, from "don't depend on one person" to "don't depend on your integrator" to "don't expect AI to fix a messy database."

On dependency and autonomy

Ricevuto repeatedly frames a well-run ERP project as one that makes the client independent, not one that locks them in. He states plainly that "Une bonne mise en place doit surtout vous rendre autonome" and that "Le but n'est pas de rester dépendant d'un prestataire pour chaque modification ou chaque question" [2]. He goes further and names the incentive problem in his own trade: "Un intégrateur a tout intérêt à ce que vous dépendiez de lui. Chaque formation, chaque nouvel utilisateur, chaque question, c'est une ligne de facture de plus" [8]. His prescribed fix is a "champion Odoo" trained internally so "la connaissance reste dans vos murs," warning that otherwise "vous ne pilotez plus votre outil, vous le louez" [8]. The same worry about single points of failure applies to people, not just vendors: "Quand votre commercial part, il ne devrait pas emporter avec lui toute la relation client" [4], and he uses a concrete test for organizational health, "Demandez à quelqu'un de retrouver un devis signé il y a 6 mois" [12].

On centralizing data

The centralization argument shows up in nearly every vertical he addresses, from field service to e-commerce. For technicians on site: "le technicien remplit sa feuille de travail directement sur sa tablette, la fait signer sur place et l'information remonte immédiatement" [5]. For e-commerce sellers juggling multiple tools: "Shopify est un excellent outil pour vendre en ligne. Le problème, c'est tout ce qu'il y a autour" [15]. And centralization is presented as the precondition for everything else he cares about, financial visibility and AI usefulness alike: "Plus vos données sont centralisées, plus vous profitez réellement des capacités de l'IA" [7].

On real-time financial visibility

He challenges founders directly on whether they actually know their numbers: "tu connais ton résultat de la semaine dernière ? Pas une estimation. Pas un ressenti. Le chiffre" [3]. The danger he names is drift into failure without noticing: "on peut glisser vers la faillite en pensant que tout va bien. Simplement parce qu'on regarde ses chiffres dans le rétroviseur" [3]. His remedy is unifying project management, invoicing and accounting so margins are visible "en temps réel pour repérer rapidement les écarts" [10], rather than discovering profitability "une fois qu'il est terminé" [10].

On AI plugged into the ERP

Ricevuto treats Odoo's native MCP server capability as a structural advantage over competitors who sell AI as an add-on: "Odoo devient lui-même un serveur MCP... Un endpoint, une clé, et votre assistant interroge vos commandes, vos stocks, vos factures," contrasting this with rivals where "L'IA posée à côté du système, jamais dedans. Ce n'est pas la même philosophie. Ce n'est pas le même budget" [13]. But he is careful to puncture hype: "Ce n'est pas magique. Si votre base est sale, l'IA vous répondra vite et mal. Si vous ouvrez l'écriture trop tôt, vous créez un problème que vous n'aviez pas" [6]. The practical case he gives is a churn-detection question, "quels clients commandaient régulièrement et ont arrêté depuis six mois," that used to cost "une demi-journée" and now takes "30 secondes" once Claude is connected to Odoo [6].

On changing the method, not just the tool

He argues that switching ERPs without changing how the business works is the most common failure mode: "Beaucoup d'entreprises changent d'ERP sans vouloir changer leur façon de travailler... c'est souvent là que le projet déraille" [9]. His conclusion is blunt: "ce n'est pas le changement d'ERP qui fait la différence. C'est le changement de méthode" [9]. The same instinct against forcing pace over adoption shows up in his repeated Big Bang critique: "La vitesse d'un projet ne se mesure pas au jour du go-live. Elle se mesure au moment où les équipes s'en servent vraiment" [1][14].

From the stage

In interviews he explains why integrators exist at all in an ecosystem that already has Odoo itself, pointing to supply and demand imbalances and the value of local proximity for SMEs and larger accounts [16]. He also describes Nalios Impact, a program extending the same digitalization work to public organizations and nonprofits (ASBLs) in Wallonia and Brussels, including VAT recovery help and a free 4500 euro project analysis, which is not mentioned anywhere in the written posts [18]. On a personal note, he mentions combat sports as a counterbalance to his daily work in software integration [16], and in the Cercle de Wallonie-Bruxelles interview he traces his path from leaving school at 17, a brief tiling apprenticeship, and supermarket work, to co-founding a 35-person company, framed around a drive to prove himself [17].

Takeaways

• Before a Big Bang go-live, consider phasing the rollout by wave; adoption speed matters more than go-live date [1][14].
• Ask whether your team can retrieve a six-month-old signed quote quickly; if not, you have a centralization problem worth fixing [12].
• Do not let sales relationships live in one person's inbox or head; centralize client history so departures do not cost you the account [4].
• Before deploying AI on your ERP data, clean the database first, a dirty base means fast, wrong answers [6].
• When choosing an integrator, ask whether they train your team to be self-sufficient or keep you paying for every question [8].
• Track project margins in real time rather than waiting for the accountant's after-the-fact report, by the time you see a loss it may be too late to fix [3][10].