Pieter Janssens

The through-line

Pieter Janssens is animated by a refusal to let anything settle. "Als ik tegen één ding niet tegen kan, dat is dat de dingen blijven zoals ze zijn. Want in mijn beleving gaan ze dan eigenlijk achteruit" [1], he says, and the shorter version he keeps reaching for is "Stilstaand water begint te stinken" [1]. That instinct runs through every part of the operation: the deliberate manufacture of sales pressure in the early years, the choice of the harder integration route after acquisitions, the insistence that AI means redesigning processes rather than buying tools, and the private test he applies to himself as a leader. The moment the list of things done well outgrows the list of things to improve is, in his telling, "de dag dat ik denk dat ik blind geworden ben voor onze eigen gebreken" [2], and a CEO in that state should be carried out the back door [2].

What has shifted is the register of ambition rather than the underlying restlessness. The famous 30% yearly organic growth iO ran from 2005 to 2020 has been abandoned; since the geopolitical turbulence of 2022 the company is only slightly positive organically, because client budgets are under pressure [1]. The energy has moved from compounding a growth number to forcing an AI-native transformation through a €220M, roughly 2,000-person business across five countries, assembled through 34 acquisitions with Waterland [1].

On what AI-native actually means

Rolling out Copilot or ChatGPT to staff is not being AI-native, and Janssens is blunt about the cost of pretending otherwise: it produces shadow AI and no ownership of your own data [1]. "AI is geen tool hè. AI is het herdefiniëren van je proces heel vaak" [1]. Being AI-native means rethinking business processes and automating them with agents and workflows, internally and in the interaction with customers [1]. He reads the market in waves: the LLM-model wave, then the compute and chips wave, and now a third wave in which the value shifts to building actual business applications, while too many companies remain stuck in an exploration phase of failing experiments [1].

His advice to entrepreneurs who feel behind is to calm down rather than accelerate. "Stop chasing the hype. Het is echt niet omdat er een nieuw lkt dat je dat moet gaan exploreren" [1]. The work is to translate AI possibilities into your own business context and to bring in an implementation partner instead of building it yourself, because home-grown solutions are precisely why so many AI innovations fail [1].

On why agencies have to know the client's business

The agencies most at risk are the ones approaching AI out of fear, bolting tools onto existing services and discovering a revenue problem underneath [1]. Janssens locates iO's advantage in a combination that is hard to fake: a business mindset that understands what the client actually does, joined to a deep technology mindset covering software architecture and AI orchestration [1]. "Ten tijden van AI moet je als agency meer dan ooit de business van uw klant kennen. Want als je de business van uw klant niet kent, kan je ook wel de shift naar AI native niet helpen oplossen voor uw klanten" [1].

On buy-and-build done the hard way

Most buy-and-build trajectories centralise shared services in a holding and let each label carry on doing its own thing; iO chose the harder route of one brand, one proposition set, one P&L and one backbone running from lead to invoice [1]. That choice makes a demand of everyone who joins, founders included: "Je moet over uw eigen schaduw kunnen stappen, want we gaan echt wel alles oprollen naar één" [1]. The argument he offers sellers is arithmetic about ownership rather than sentiment: "Je hebt misschien beter inderdaad een kleiner deel van een groter geheel dan een groter deel van een kleiner geheel" [1]. It follows that the long-term plan and the post-acquisition transformation belong in the very first conversation, because surprising founders after the deal is signed produces misery [1].

Diligence, likewise, has to go a level below the headline multiple of EBITDA. He looks at client type fit, client concentration risk, recurring revenue, human capital loyalty and whether people are paid fairly, the depth of middle management, and whether organic growth matches or beats market growth [1]. The same logic shaped the deal with Waterland, where a deliberately non-traditional 50-50 equity split meant the founders never mentally felt they had sold the company or handed over the keys, preserving the autonomy needed for the next growth wave [1].

On organic growth as the honest signal

Acquisitions are an accelerator and nothing more; organic growth has to be the main engine [1]. The diagnostic is unforgiving: if all your growth comes from acquisitions, something is fundamentally wrong with your proposition or your culture [1]. The structure is built to serve that engine directly, organised by capability rather than by location, with one new-business team and capability teams spread across Herentals, Antwerp, Ghent and Brussels, and 98% of revenue coming from clients whose decision centre sits in a single country [1].

On putting yourself on the edge of the abyss

The original growth mechanism was psychological. "Ik hou ervan om mezelf zo net op de rand van de afgrond te zetten" [1]. In practice that meant hiring a new employee precisely when the pipeline was comfortably full, which recreated sales pressure and forced him back out to hunt for work for more people [1]. It is the same reflex as the stagnant-water line, applied to his own comfort rather than the company's.

On hiring people who are ten times better

Complementarity is the minimum bar, not the goal. "Je moet ook proberen te zoeken naar mensen die niet alleen complementair zijn maar in een bepaald domein gewoon 10 keer beter zijn dan ikzelf" [1].

On two time horizons at once

He holds a very long view and a very short one simultaneously, and pushes the organisation toward the short one on purpose: "Je hebt een heel lange termijnblik en een heel korte termijnblik. En ik probeer ook iedereen te focussen op het volgende kwartaal om vooral die actie mindset in het bedrijf te hebben" [1].

On boundaries you defend like targets

The restlessness stops at the front door, and only because it is fenced in with rules as concrete as any business objective: no weekend work, laptop closed at 4pm on Friday, a minimum of three dinners a week at 6pm, and bringing the children to school once a week [2]. These are scheduled with the same strictness as commercial targets [2].

Takeaways

• Treat AI as process redesign, not procurement; distributing Copilot or ChatGPT without rethinking workflows creates shadow AI and leaves you without ownership of your data [1].
• Ignore each new model release and spend the effort translating AI into your specific business context, working with an implementation partner rather than building in-house [1].
• In buy-and-build, put the long-term plan and the post-deal transformation on the table in the first conversation; founders who are surprised afterwards make everyone miserable [1].
• Look past the EBITDA multiple in diligence: client fit and concentration, recurring revenue, staff loyalty and pay fairness, middle management depth, and organic growth versus market growth [1].
• Use organic growth as your honesty check; growth that comes entirely from acquisitions signals a broken proposition or culture [1].
• Manufacture your own pressure by hiring when the pipeline feels comfortable, so you are forced back into selling [1].
• Hire for people who are ten times better than you in a given domain, not merely complementary [1].
• Watch for the day your list of things done well is longer than your list of things to improve; that is the day you have gone blind to the company's flaws [2].