The full version of the study is available exclusively to AFRC members; see the link at the end of the article.
Artificial intelligence is fundamentally transforming customer relations.
For several years, conversational AI has been treated as a testing ground: exploring use cases, improving the user experience, and demonstrating the technology’s potential. That phase is coming to an end.
Today, the topic takes a different turn.
Scaling up is the real challenge.
This is precisely what the study conducted by AFRC and FROG (Capgemini)—to which Zaion contributed alongside numerous customer relationship management stakeholders—highlights.
His conclusion is clear: AI is entering a phase of industrialization.
Scaling Up AI: New Challenges for Customer Relations
The widespread adoption of conversational AI marks a turning point for businesses.
While pilot projects allowed for some flexibility, production deployments impose much stricter requirements.
On a large scale, the challenges take on a different character. Volumes increase, operating costs become apparent, and compliance requirements become more stringent.
Companies must therefore grapple with a more complex challenge: controlling the costs associated with using models, ensuring the compliance of model outputs—particularly in regulated sectors—ensuring the traceability of decisions, and maintaining consistent quality over time.
Industrializing AI is therefore not simply a matter of scaling up. It requires a fundamental rethinking of architectures, processes, and governance.
Conversational AI: Why Mastery Is Key
One of the study’s key findings is clear: the value of AI no longer lies solely in the technology itself, but in its operational mastery.
The rise of generative AI has led to significant advances in language understanding and conversational fluency.
But this shift brings with it new challenges: greater variability in responses that is harder to control, increased governance and audit challenges, as well as risks in the event of misuse or abuse.
On a large scale, AI must do more than just convince people.
It must operate in a reliable, measurable, and well-governed manner.
Hybrid architectures: a standard for scaling up AI
The AFRC x FROG study highlights a clear trend: the emergence of hybrid architectures for deploying conversational AI at scale.
Today, 100% generative approaches are showing their limitations in demanding environments.
Conversely, the most robust systems rely on sophisticated orchestration, combining deterministic logic for critical steps—such as authentication, transactions, or compliance—generative logic where understanding and fluidity add real value, and comprehensive orchestration that enables the management, monitoring, and measurement of user journeys.
The issue is no longer about choosing a technology; it is about building orchestration capabilities.
Transforming Customer Relationships: A Business Challenge
Scaling up AI in customer relations is no longer just a matter of technology.
This is now a business issue.
The customer experience, customer relations, and operations departments play a central role in prioritizing use cases, balancing performance, cost, and risk, and integrating AI solutions into existing processes.
It is this integration of technology and expertise that determines the success of large-scale projects.
Feedback: Zaion's Approach
As part of this study, Zaion shared feedback from production deployments, particularly in demanding sectors such as banking, insurance, and public services.
These environments highlight a simple operational reality: a chatbot’s performance isn’t measured solely by how smoothly it interacts.
It depends just as much on its ability to comply with strict regulatory requirements, integrate into complex business processes, and deliver measurable results over the long term.
This feedback from the field confirms a belief: conversational AI only creates value when it is designed as a controlled system.
The Industrialization of AI: A Turning Point for Businesses
Conversational AI is now entering a phase of maturity.
Its industrialization entails higher standards, greater project structuring, and the ability to generate a measurable impact on operational performance.
The companies that will succeed in scaling up will not be those that adopt technologies the fastest, but those that know how to integrate them intelligently into their operations.
Above all, those that strike the right balance between intelligence, control, and reliability.
