What our study of 11 French CX leaders reveals
Where exactly should the line sit between AI and human by 2030? That's the question we put to 11 CX directors at LVMH, Carrefour, Engie, SFR, MACIF, Volkswagen, La Banque Postale, Matmut, Solocal, Mutuaide, and Vattenfall. Their answers converge on a paradox few organisations have anticipated.
This story illustrates a paradox our CX Horizon 2030 study uncovered: at the very moment generative artificial intelligence makes it possible to automate almost the entirety of a customer relationship, that's precisely when the brands that perform best are reinvesting in humans.
This article answers one simple question: where exactly should the line sit between AI and human by 2030? You'll find what the field says, not the theory, on four specific points: why customers are becoming less and less patient, why human contact is becoming a rare competitive advantage, exactly where AI genuinely creates value without damaging trust, and how the customer relations business model needs to evolve to keep up.
Acceptable delay no longer exists. That's not a figure of speech, it's a behavioural shift documented by every director interviewed in the study.
The customer's benchmark is no longer their own sector. A policyholder no longer compares their insurer to another insurer, they compare it to Amazon. An energy brand is no longer judged against its direct competitors, but against Uber. This shift of the speed standard from the on-demand economy into every sector, including those with long decision cycles, creates an irreversible ratchet effect: once the bar is set at instant, it never comes back down.
Ludovic Le Pape, Head of Development at Mutuaide, illustrates this shift with a simple situation. A customer broken down on the roadside doesn't want to know that a tow truck is coming. They want to know if it takes 35 minutes or 52. The difference isn't in the service delivered, it's in the precision of real-time information.
This pressure has a second, more insidious effect: complexity no longer resolves in self-service. Many organisations pushed customers toward digital journeys to cut costs. The problem is that strategy works for simple tasks and collapses the moment things get complicated. A change to an insurance policy, a tax question about an electric vehicle, an unreadable electricity bill: these are topics customers can no longer handle alone, and that digital alone fails to clarify.
Here's what most cost-cutting plans overlook: forcing digital on a complex topic doesn't remove the call, it delays and worsens it. The study names this phenomenon Failure Demand, demand generated by the failure of the digital journey. A customer who fails three times in an online journey before calling arrives frustrated, with a problem that's now harder to solve. The call that was supposed to cost less ends up costing more than if it had happened directly.
now expect more personal contact than before. In a world saturated with generative AI, authentic human conversation becomes a scarce resource, and therefore strategically valuable.
It's in this context that the study's most-cited figure takes on its full meaning: 81% of customers now expect more personal contact than before. The consequence of this tension goes further than a simple satisfaction number.
Here's the paradox most CX strategies miss: the more generative AI produces content, the more it produces it for free, at infinite scale, the rarer human conversation becomes. And a scarce resource always obeys the same economic law: it gains value. That's exactly what's happening in customer relations, and it's the opposite of what many leadership teams anticipated back in 2020.
Carole Brion, Head of Customer Relations at Carrefour, draws a sharp cultural observation comparing the French market to the UK market. Across the Channel, automation dominates customer relations. In France, customers retain a visceral attachment to interaction. They need to be able to express themselves, exchange, be heard. Even in a digitised world, 80% of French customers still prefer going through a traditional checkout, if only for the pleasure of a hello, a goodbye, a glance, even at the cost of some pure efficiency.
In a retail sector where price wars crush every margin for manoeuvre, this attachment becomes a lever that neither price nor product can offer anymore: the last real differentiator left.
We're heavily investing in the customer experience. It's precisely what sets us apart: we're strengthening the bond of trust with the brand.
Hugues Beaujouan, Head of Customer Services at LVMH, has theorised this principle under the name clienteling: a proactive, personalised relationship where the advisor knows the customer's history to make the most relevant offer. The ambition for 2030 isn't to reserve this treatment for premium customers alone, but to democratise it: offer luxury-level attention to mass-market customers, thanks to data that gives the advisor a perfect memory of the relationship.
This shift concretely transforms the role. The 2030 agent won't be evaluated on their ability to recite a script, a task a voice AI already performs better, without fatigue and without an accent. They'll be evaluated on their ability to depart from it to adapt to the real emotional state of the person on the other end. It's a complete reversal of recruitment and evaluation criteria.
None of the 11 leaders interviewed envisions a 100% automated customer service by 2030. That's the exact opposite of the prevailing discourse around autonomous agents and next-generation chatbots.
Dominique Russo, Head of Customer Experience at MACIF, sets the limit precisely: the sincerity of a relationship is incompatible with a robot pretending to be human. And Thierry Suquet, Head of Customer Experience at Volkswagen, adds an often-overlooked economic dimension: AI has no room for error. A human advisor can apologise, recover, turn an incident into a moment of trust. A bot that hallucinates information causes a break that nothing repairs. The cost of the error isn't the same.
Where AI creates immediate value is in the back office. Post-call, it listens, transcribes, and automatically categorises the conversation in the CRM. This recovered time, between 15 and 25% of total handling time, isn't reinvested in headcount cuts. It's reinvested in the relationship. The concept of the augmented advisor rests on this principle: freeing up cognitive time for the relational dimension, the one no script and no bot can reproduce.
The augmented advisor: technology in service of the relational dimension, not in its place.
What remains is a grey area the study names directly: the uncanny valley of trust. 60% of managers don't yet trust AI to interact alone and unsupervised with a customer. This caution validates a strategy embraced by nearly the entire Armatis panel: AI in the back office, without reservation. AI in the front office, yes, but under human control.
In the old world, the provider benefited from the phone ringing. More calls, more billed hours, more revenue. In the 2030 world, it's the opposite: the brand wants the phone to stop ringing. Fewer calls means less friction, lower costs, higher satisfaction. Jean-Rémy Dudragne, Head of Customer Experience at Engie, sums up this shift in one sentence: the interests of the provider and those of the brand, long aligned, are becoming completely desynchronised.
The historic BPO model, built on selling hours or processed transactions, is therefore structurally challenged, not by falling demand, but by the very efficiency of the AI it deploys. This misalignment of interests remains baked into most current contracts, which continue to index pay on volume rather than resolution.
The market doesn't yet have a fully mature alternative model, but the direction is clear: toward hybrid pay, indexed on performance and value created rather than hours consumed. It's a contractual paradigm shift that affects brands and their providers alike, Armatis included.
Artificial intelligence is also redrawing the map of contact centres, but not in the direction you'd instinctively expect. For the first time in the history of remote customer relations, the language constraint is dissolving thanks to real-time translation and accent neutralisation. But the cultural constraint is strengthening.
Dominique Russo (MACIF) sums up this tension with a sharp formula: consumers expect brands to be honest and committed. There isn't, on one side, corporate values, and on the other, a customer-oriented posture. A local presence that contributes to regional development is a strong marker of closeness and trust.
Cultural closeness, an advantage machine translation doesn't reproduce.
The model emerging from the study has a name: smart-shoring. It's no longer about stacking destinations purely by cost (inshore, nearshore, offshore), but matching the nature of the interaction to the culture best suited to handle it. Crisis management, strong emotion, premium cases: that's the territory of local closeness, a form of relational haute couture. Standardised transactional processing, meanwhile, continues to benefit from offshore expertise hubs that have proven themselves over twenty years.
Technology doesn't solve everything: while the machine translates words, it doesn't translate implicit cultural codes. LVMH and La Banque Postale insist on this point: the customer relationship is an exchange of rapport. Humour, references to current events, a fine understanding of a life context like the healthcare system, often escape an offshored interaction, even one assisted by a very capable AI.
The question this article raises isn't "should we use AI?" That question is already settled. The real question, the one the 11 directors interviewed put to their teams, is: for which conversation are you irreplaceable?
Because that's where everything plays out. The full study goes further on this point with the four pillars of CX performance, a strategic positioning tool (premium, leader, or volume), and all 11 interviews in full. It answers this question not in theory, but based on what LVMH, Carrefour, Engie, SFR, MACIF, and six other CX leadership teams are doing right now.
Contact our teams to discuss your challenges and find out how we can support you
Join the leaders who trust our multilingual and technological expertise.