As the current President of SP2C (Professional Union of Contact Centers) since July 2022, Caroline Adam continues with determination her second term at the head of this representative organization for a key sector of the French economy. Caroline Adam has successfully structured a collective strategy to promote the value, professionalization, and influence of the sector. She embodies a recognized capacity for influence, working closely with political, economic, and social institutions. As an experienced lobbyist, she advocates the sector’s priorities to public decision-makers while supporting her members through structural and regulatory transformations of the profession.

In this interview, she discusses the sector, its changes, and the role of AI in the evolution of these professions.

SP2C, a strategic federation

SP2C, the Professional Union of Contact Centers, operates as a true sectoral federation affiliated with MEDEF. It brings together nearly 75% of the French outsourced customer relations market, consisting of nine major companies covering all channels of interaction between brands and users: telephone, email, chat, social media, or mail.

Behind consumer brands, these major BPO (Business Process Outsourcing) players operate behind the scenes while being crucial: €3.56 billion in revenue, over 100,000 employees in France. An industry that even exceeds aerospace in employment volume. “The French are world champions in customer relations,” recalls Caroline Adam, citing sector leaders such as Teleperformance, Concentrix (formerly Webhelp), or Foundever (formerly Sitel).

A progressive integration of AI

Digitalization has transformed practices for over a decade. AI, meanwhile, is part of a technological continuum, with notable acceleration over the past three years. The approach remains pragmatic: no massive replacement of humans, but rather strengthening skills and productivity.

Among pioneering uses, Caroline Adam cites predictive AI, deployed to anticipate call flows based on activity peaks (holidays, sporting events, government announcements), absenteeism, or seasonality. The objective: optimize workforce mobilization in real time.

Another emblematic example: “AI ergonomists.” Previously, a human ergonomist observed agents’ gestures, clicks, page openings to recommend improvements. Today, AI analyzes thousands of sessions and proposes optimization scenarios subsequently validated by IT managers. This form of algorithmic ergonomics foreshadows a hybridization of professions.

AI in service of emotion and semantics

In stressful or emergency situations, such as during the health crisis, AI-powered vocal bots automatically responded to simple questions (vaccination, schedules, nearest center). They redirected calls detected as anxious through tone and timbre analysis to a human agent.

On the semantic analysis side, use cases involve deciphering customer feedback at scale. Striking example: in delivery services, if 80% of complaints mention delays, this isn’t always what generates major dissatisfaction. Rather, it’s the nature of the product and the emotional moment (she cites the example of the wedding dress delivered late or non-conforming) that crystallize customer friction. AI enables fine prioritization of corrective actions here.

New professions thanks to generative AI

Generative AI uses are multiplying. On the translation side, they enable maintaining 24/7 service quality for demanding international customers, as with BlueLink, which manages luxury brands. AI translates customer requests, allowing a French-speaking agent to respond in Dutch or Japanese via email.

In training, AI plays a dual role: it learns from human advisors (“coach the bot”) and serves as a virtual trainer for newcomers. It simulates customer cases, offers corrections, and tracks performance. The old supervision model—listening to 10 calls per month—is replaced by automated analysis of 100% of interactions. The training plan is thus personalized with great precision.

Another example: in insurance, during a claim, AI automatically transcribes exchanges, enriches the CRM, suggests relevant questions, and schedules the appointment with the expert. It allows the agent to focus on emotional aspects and human relationships.

AI agents and agent experience: a new frontier

As conversational agents multiply, the question of their integration into digital journeys arises. Caroline Adam emphasizes the need for compatibility between the AI agent and the digital interface of the targeted service. “If the travel agency’s website, for example, is not designed to cooperate with an AI agent, the booking will fail.”

But beyond the B2C use case, she warns about the ethical and strategic implications of automating B2B negotiation processes. Replacing human buyers with AI agents configured to negotiate within a predefined range could ultimately lead to reductive compromises, erasing long-term vision or relational subtleties. AI would then be too “literal” a tool for such a contextual and human challenge as complex negotiation.

RPA, open banking, and extreme automation cases

Some innovations combine AI and automation (Robotic Process Automation). In e-commerce, AI-powered RPAs consult various logistics platforms in real time to inform a customer about the status of a multi-product order.

Another spectacular case: Smart Push, a French startup that allows, with user consent, to verify in real time the veracity of a credit file via open banking data. This time saving drastically reduces document fraud.

A progressive acculturation, a systemic vision

SP2C helps its members structure pragmatic AI acculturation. The advice: start from the customer journey, identify pain points, test with POCs, and don’t give in to trends. “It’s not because you’re doing AI that it’s useful. The essential thing is to measure ROI and provide real value.”

This test-and-learn method also helps limit failures: “80% of POCs launched by large companies have failed,” she reminds.

Controlled HR impacts and new roles

The observed productivity gains (15 to 30%) do not translate into layoffs but into reduced recruitment. Certain profiles are emerging, without engineering training: “data stewards,” or database managers, who ensure the quality of information feeding AIs.

An experienced customer advisor can evolve toward these data professions. The challenge is to maintain employability in a sector known for its turnover while moving upmarket.

Toward an industrial AI ethics?

Faced with risks of dehumanization or algorithmic hallucinations (for example, a chatbot that insults its own company or proposes absurd rates), ethical reflection is imperative.

La Poste has already published a responsible AI usage charter, integrating its service providers. For now, its members comply with brand charters and European regulations (AI Act, DORA, NIS2).

A systemic and human vision

Finally, Caroline Adam recalls the principle of symmetry of attention: improving customer experience also requires caring for that of employees. AI must serve to relieve agents of tedious tasks to refocus them on listening, empathy, and personalization.

“AI is not an end in itself. It’s a lever to be used with rigor, ethics, and strategy,” she concludes. An embodied, lucid, and demanding voice for a sector that combines technology, public utility, and human skills.