
Steven Van Belleghem
Steven Van Belleghem co-founded Nexxworks in 2015 alongside Peter Hinssen and Rik Vera, positioning himself as the customer experience and AI voice within the trio. He is based in the Ghent area (Aalter) and holds a professor position at Vlerick Business School.
He has delivered more than 1,000 keynotes across 40+ countries, making him one of the most traveled Belgian speakers globally. His book "A Diamond in the Rough," published in 2024, was ranked the Forbes #1 must-read business book of that year.
Steven's content focus sits at the intersection of AI adoption and customer-centricity, a position that resonates strongly with enterprise audiences navigating Digital Transformation.
Insights & takeaways
Steven Van Belleghem's thinking across these sources converges on one central bet: artificial intelligence is dissolving the old rules of customer experience, and the companies that win will be the ones that understand what becomes commodity and what becomes precious. His clearest formulation of this is blunt: "De core van wat dat AI doet in customer service is goede service veranderen in een commodity" 8. Good, frictionless service used to be a differentiator because it was scarce; he illustrates this with the photography analogy, comparing the shift to when unlimited digital photos made a single picture worthless compared to the era of the 12-exposure film roll . His conclusion is that "Good customer service will no longer be a differentiator. It is becoming hygiene" , which pushes the real competitive battleground toward something he calls "deep loyalty" — a theme significant enough that it becomes the title of his forthcoming book 8.
A second major thread is his framework for agentic AI, which he is careful to distinguish sharply from chatbots. "Agentic AI is een AI tool die in jouw naam acties kan uitvoeren" 910 — it doesn't just talk, it acts, buys, books, and pays. He predicts this will restructure digital life within a decade: "Ik zie binnen 10 jaar de rol van apps vervangen door Agents" 910, with people issuing commands to personal agents the way they would to an employee. Tied to this is his "99/1" model of human contact: 99% of interactions just need efficiency and can be automated, while the remaining 1% — complex, emotional, high-stakes — is where human presence still wins trust, meaning "de frequentie van menselijk contact... gaat enorm zakken, maar die paar momenten dat het nog menselijk is gaan enorm in belang toenemen" 910. He backs this with a striking reversal in consumer sentiment: where people once rejected programmed empathy, "Vandaag zijn wij allemaal klaar, zeggen wij van: geef ons maar een vriendelijke computer boven een onvriendelijke mens" 910.
Van Belleghem is consistently skeptical of hype-driven timelines, even as he insists on the scale of the transformation. "Het is niet zo moeilijk om de toekomst te voorspellen. Het is heel moeilijk om de timing ervan te voorspellen" 910 is a recurring caution: technical capability has existed for automating most customer service for two years, yet adoption lags because "organizations need time and people to implement" 910, and he ties adoption speed directly to whether a CEO personally uses and champions the technology 910. This is paired with a real-world observation from the LLM usage data he discusses on his RADAR podcast: "companies talk only about efficiency in their AI discussion, customers mainly go for emotional use cases" — a paradox that reinforces his broader argument that organizations misjudge what customers actually want from AI.
Underneath the AI commentary sits a well-developed philosophy of customer-centric culture that predates and now intersects with his AI thinking. His "95/5 rule" holds that companies overbuild restrictive policies to guard against the unreasonable 5% of customers, punishing the reasonable 95% in the process 8 — captured in his observation that "Een klant verwacht geen perfectie, hè. Een klant verwacht positieve intentie" 8. He argues cultural change is driven overwhelmingly by leadership behavior, not policy documents, and that incremental "friction hunting" beats grand transformation projects because it's cheap, fast, and builds a compounding culture of attentiveness 8. He's also candid about organizational courage, or its absence: "Op het moment dat je twijfelt, twijfel je eigenlijk niet meer. Dan ontbreekt je gewoon de durf om de volgende stap te zetten" 8, and he plays devil's advocate deliberately to expose when an idea is being killed by risk-aversion rather than substance: "Ik ga eventjes de advocate of the devil zijn. Als iemand dat zegt, dan weet je: het is game over voor het idee" 8.
His view of employees is inseparable from his view of customers: "Klanten kunnen niet gelukkig zijn als medewerkers ongelukkig zijn" 8, and he frames the willingness to accept short-term pain for long-term trust as a leadership test — "Ben je bereid jezelf pijn te doen op de korte termijn om vertrouwen te kweken op de lange termijn? Customer first without compromise" 8. He extends this into concrete organizational advice: large companies need both "Michelin chefs" who run reliable process and "street food chefs" who improvise and innovate 8, and he warns against blanket instructions like telling staff to "say no more often," which he says installs a race to the bottom rather than building genuine judgment 8.
Notably, Van Belleghem frames his own career, and by extension his credibility as a forecaster, with some humility: "Als ik kijk naar mijn carrière, dat is allemaal eigenlijk een opeenstapeling van toevalligheden. Alles is toevallig gelopen. Het is niks gepland geweest" 8. That mix of conviction about big structural shifts and modesty about prediction and planning is characteristic of how he presents ideas throughout — confident on direction (agents replacing apps, service becoming commodity, loyalty becoming the new differentiator), cautious on exact timing and on claiming any grand personal design behind it.
The practical takeaways that recur across his output are consistent: treat AI-driven service automation as inevitable and near-term but organizationally slow: invest deliberately in the small number of human touchpoints that will carry outsized weight; hunt friction continuously rather than launching one big transformation; and diagnose your own culture by asking whether leadership's micro-decisions actually protect customers under pressure, since that, more than any stated policy, decides whether customer-centricity is real. His three 2026 books — a youth thriller, "Deep Loyalty" on customer experience and AI, and a practical culture book built substantially from new cases beyond his earlier "De Ruwe Diamant" — reflect this same throughline: translating structural shifts in AI and commoditized service into very specific, applicable guidance for how companies and their people should behave.
- The 95/5 rule: max 5% of customers are unreasonable, but our brains overweight them, so companies build protective rules and procedures that punish the 95% of normal customers — customer-centricity improves when you mentally accept the 5% and design for the 95%.
- A large company cannot fully operate like a startup; you need both 'Michelin chefs' (process-driven operators running the machine) and 'street food chefs' (improvisers) to combine reliable delivery with innovation.
- Culture change is driven 70-80% by leadership: the micro-communications and micro-decisions leaders make determine whether a customer-centric culture takes hold.
- Cumulative 'friction hunting' — removing hundreds of small frictions per year — beats one big transformation project: it costs little, has immediate effect, and installs a win-after-win culture of customer attentiveness.
- AI is turning good, frictionless customer service into a commodity/hygiene factor; once everyone achieves friction-free, it stops being a differentiator, and the next battleground becomes 'deep loyalty' — customers feeling proud to belong to your brand.
- Employees will always choose an angry customer over an angry boss, because the customer leaves and the boss stays — so unless leaders visibly prioritize the customer relationship in the first crisis moment, employees optimize for pleasing the boss forever after.
- Be 'micro ROI negative': individual acts of generosity (Taylor Swift's personalized gifts to 50 fans, Stanley replacing a burned car) look like costs but create leveraged word-of-mouth; companies kill these ideas via meetings and 'devil's advocate' risk extrapolation.
- The elevator attendant vs. doorman framework for AI: automate roles that only push buttons, but keep and even invest more in 'doormen' whose human presence adds welcome, safety and advice — companies that automate both will regret it, as Klarna did.
- Startups embed customer-centricity by using their own product relentlessly and spending deep time with early customers — like Airbnb's Brian Chesky sleeping months per year in listings on his own platform.
- Telling teams to 'say no more often to customers' installs a negative race to the bottom; instead build the mindset and capacity to say yes more than today, without saying yes to everything.
- Customer-centric doesn't mean free: people will pay more for good service; extra services can be paid — offering them at all is already customer-centric.
- During crises, marketing budgets get cut first, but that's when investing stands out most — many major companies (Airbnb, Uber) were born out of the 2008 crisis.
- Agentic AI is fundamentally different from chatbots: it executes actions in your name (selecting, buying, paying with your credit card) rather than just communicating — the 'ChatGPT moment' for agents hasn't happened yet but will in 2025-2027.
- Within 10 years apps could be replaced by agents: a financial agent instead of a banking app, a travel booker instead of booking.com, driven by voice commands like you'd give an employee.
- Consumer preference has flipped in 10 years: people used to choose 'no empathy' over programmed empathy; today they prefer a friendly computer over an unfriendly human, and studies show consumers often find ChatGPT more empathetic than human agents.
- In 99% of service interactions customers only want efficiency; the 1% complex, emotional, high-stakes moments require humans — the frequency of human contact will plummet but its importance will soar, meaning all traditional service-center KPIs and training should be thrown out.
- AI job fear mirrors past fears: women entering the workforce and the internet didn't destroy net employment but added economic power and new jobs; AI will do the same via productivity gains and new company formation.
- Adoption will be slower than Silicon Valley thinks: automating customer service has been technically possible for 2 years yet almost no company has done it at scale — company speed largely depends on whether the CEO personally uses the technology, just like Facebook advertising in the 2000s.
- Toyota building Woven City is a prime example of Day After Tomorrow thinking: one of the world's biggest carmakers builds a futuristic city with no room for classic cars, as a playground to test mobility innovations before exporting them to Tokyo or Osaka.
- In San Francisco many women prefer driverless Waymos over taxis with a strange driver — new safety perceptions and ways of thinking emerge around autonomous vehicles.
- Purpose is not necessarily linked to work: 100 years of automation from dishwashers to ChatGPT has shifted time from things we must do to things we want to do, and AI is the next phase of that.
- Agentic AI differs from chatbots because it executes actions in your name (buying, booking, paying with your credit card), not just communicating — the 'ChatGPT moment' for acting agents hasn't happened yet but will in 2025-2027.
- Within ten years apps could be replaced by personal agents: a financial agent instead of a banking app, a travel booker instead of booking.com — you give commands like you would to an employee.
- Consumer attitudes have completely reversed in ten years: when 'When Digital Becomes Human' came out people chose no empathy over programmed empathy; today everyone prefers a friendly computer over an unfriendly human.
- The 99/1 framework: 99% of customer interactions just require efficiency and can be automated, but the 1% that are complex, emotional or high-stakes is where human contact builds trust and wins the business — so frequency of human contact will plummet while its value soars, invalidating traditional service-center KPIs.
- AI job fears mirror past shifts: women entering the workforce and the internet added economic capacity and new jobs rather than one-for-one displacement; AI will bring a productivity leap and enable tiny teams to build large companies.
- Automating nearly all customer service has been technically possible for two years, yet almost no company has taken big steps — real-world adoption is far slower than Silicon Valley timelines because organizations need time and people to implement.
- Company AI adoption speed is largely determined by whether the CEO personally uses and loves the technology — the same pattern seen with Facebook advertising in the 2000s.
- Toyota building Woven City shows one of the world's largest carmakers designing a future city with no room for the classic car — a flagship example of Day After Tomorrow thinking with real innovation budget behind it.
Career
Intracto#117factoryPartner and Strategic AdvisorSep 2018 – Present
- Plan International BelgiumBoard MemberAug 2018 – Present
- nexxworksCo-founder and Board memberMay 2015 – Present
- B-ConversationalKeynote speaker and AuthorSep 2012 – Present
Vlerick Leuven Gent Management School#375factoryProfessorOct 2010 – Present
Hello Customer#372factoryBoard MemberApr 2018 – Mar 2021
- SnackbytesCo-Founder and Board memberJan 2015 – Sep 2018
- Zembro#205factoryCo FounderMar 2015 – Mar 2017
- InSites Consulting#181factoryAuthor of "The Conversation Company'Jan 2012 – Sep 2012
- InSites Consulting#181factoryAuthor of "The Conversation Manager"Jan 2010 – Sep 2012
- InSites Consulting#181factoryManaging PartnerJul 2008 – Sep 2012
- InSites Consulting#181factoryDirector Brand & Conversation ResearchJan 2007 – Jun 2010
- InSitesSenior ConsultantMar 2001 – Dec 2006
Vlerick Leuven Gent Management School#375factoryResearch assistantJul 2000 – Mar 2001
Ghent University#2schoolApplied economics, Marketing1995 - 2000
- OLVA#211school
- University of California, Berkeley, Haas School of Business#284schoolMarketing & High tech marketing
From public career histories · 17 entries
Media & appearances
3- 8podcastBen's Mentors · 15 Jan 2026
Steven Van Belleghem shares his career journey and customer-centricity philosophy: the 95/5 rule, friction hunting, empowering employees, being micro-ROI negative, and why AI will commoditize frictionless service, making 'deep loyalty' and community the next differentiator.
- 9podcastVirtual · 05 Feb 2025
Steven Van Belleghem joins the Virtual podcast to explain agentic AI — AI that acts on your behalf instead of just chatting — predicting agents will replace apps, enable one-person unicorns, and shrink but intensify human customer contact (his '99/1 moments' framework).
- 10podcastVirtual · 15 Jan 2025
Steven Van Belleghem joins the Virtual podcast to explain agentic AI — AI that acts on your behalf — predicting agents will replace apps within 10 years, human contact will drop in frequency but rise in value, and AI will boost rather than destroy jobs.