A major ERP migration happens once every fifteen to twenty years, at the same moment the Dutch AI Act is moving from legislation into enforcement. For Dutch enterprises, that overlap makes preparation the whole game. Treating a migration as a problem to patch, so the business ends up where it started but on a newer, supported system, wastes the one structural opportunity to rebuild the data foundation an organisation gets for a generation. Writing in Computable.nl, our Senior Value Partner Rad Parvin argued that migrations deserve the same preparation a football club gives pre-season.
Rad's argument in Computable.nl
Rad's piece, published on 12 August, argues that large ERP migrations get planned like a problem to be fixed rather than treated as a rare structural opportunity. Most programmes configure the new system to match the old one, load the data across, and only check afterwards whether the result answers the right business questions. By then, the lineup and tactics are already set.
“Large ERP migrations deserve the same lens as pre-season football training, but are almost always approached as a problem to be fixed. The goal is to end up where you started, just on a new, supported system. That framing is wrong.” – Rad Parvin, Senior Value Partner, Valliance. Translated from Dutch, Computable.nl, 12 August 2026.
Why the timing matters for Dutch enterprises
The timing is not coincidental. The Dutch government published its draft AI Act Implementation Act in April 2026, setting out a decentralised supervisory model across eight sectoral regulators. In its March 2026 report, the Dutch Data Protection Authority, the coordinating supervisor for algorithms and AI, warned that organisational preparation for the incoming AI Regulation is lagging behind the pace of adoption.
An ERP migration is exactly the moment those two pressures meet. It is the one point in a fifteen-to-twenty-year cycle when an enterprise rebuilds its core data model, arriving just as regulators start asking harder questions about how that data and its AI use are governed. Get the data model right during the migration, and the AI Act's governance and traceability requirements become easier to satisfy by design, rather than retrofitted later.
What this means for how a migration gets planned
Rad's central point in the original piece is to plan the questions the business needs answered before choosing the system, not after. A well-designed ontology, a live model of what the organisation runs on and how those things relate, combined with AI, gives leadership a flexible way to keep reading the game as conditions change, rather than locking in a lineup before kick-off.
The full mechanics of that approach are set out in Pre-season thinking for ERP migrations. This note exists to flag why it matters more, right now, for anyone running a Dutch migration alongside AI Act preparation.
Read Rad Parvin's full piece, Een verloren seizoen begint met slechte voorbereiding, ook bij erp-migratie (in Dutch), on Computable.nl
FAQs
How often do large-scale ERP migrations happen?
Once every fifteen to twenty years, according to Rad Parvin's piece in Computable.nl. That makes each one a rare chance to rebuild an organisation's data foundation, not just to swap systems.
What happens if a migration just does lift-and-shift?
The old estate gets transferred one-to-one onto a new, supported system, but the data fragmentation and workarounds move with it. The opportunity gets postponed to the next migration, which, on a fifteen-to-twenty-year cycle, is a long wait.
What is an ontology, in this context?
A living model of an organisation, describing what it runs on (orders, products, factories, customers), how those things relate, and what actions are possible on them. ERP systems record transactions. The ontology gives those transactions meaning, so people and AI can act on them together.
What is pre-season thinking for ERP migrations?
It is the argument that a major ERP migration should be planned like a football club's pre-season: deciding what questions the business needs to answer, and what data model supports that, before the system itself is chosen. Most migrations do this backwards, configuring the system first and checking the data model afterwards.
Why does the Dutch AI Act make ERP migration timing more urgent?
The Dutch government's draft AI Act Implementation Act, published in April 2026, sets a decentralised supervisory model across eight sectoral regulators. The Dutch Data Protection Authority's March 2026 report found that organisational preparation is lagging behind AI adoption. Enterprises migrating ERP systems now have the chance to build governance and data traceability into the new system from the start, rather than retrofitting it once a regulator starts asking questions.
