Customer ServiceAnalysis

Humanized AI: Why the Balance Between Automation and Empathy Defines Success

Analysis of how companies like Carrefour, Boticário, and Synthesia found the ideal balance between automated efficiency and human empathy in customer experience.

29 de novembro de 2025

9 min read

Source: itforum.com.br

+33pp

CSAT Synthesia

3

Carrefour Personas

NPS 90

Magalu

The most common misconception about AI in customer service is treating it as a binary choice: either full automation or human service. The most successful cases demonstrate that the real differentiator is the balance — knowing when to automate, when to humanize, and how to create AI that is itself perceived as humanized.

This analysis examines three distinct approaches to this balance — Carrefour (AI personas), Boticário (intelligent personalization), and Synthesia (self-service with smart escalation) — to extract principles for designing customer experiences that are simultaneously efficient and empathetic.

+33pp

CSAT Synthesia

3

Carrefour Personas

NPS 90

Magalu

AI Persona Design

The Carrefour Approach: AI Personality as Brand Strategy

Carrefour's most innovative decision wasn't implementing AI — it was creating three distinct personas (Carina, João, Samy) that reflect the identity of each brand.

Design Principles: - Carina (Carrefour): friendly, practical, solution-oriented — reflects the everyday convenience brand - João (Atacadão): direct, objective, value-focused — reflects the cash-and-carry warehouse model - Samy (Sam's Club): sophisticated, exclusive, experience-focused — reflects the premium membership club

This approach recognizes that humanization isn't about pretending the AI is human — it's about the AI communicating in a way that's authentic to the brand. A customer interacting with Carina gets a different (and appropriate) experience than one interacting with Samy.

The critical complement is intelligent escalation: the AI identifies signals of frustration, complexity, or emotional sensitivity and makes a seamless transition to a human agent who receives the complete context of the conversation.

The lesson is: AI persona design is as strategic as visual brand identity design.

When to Escalate to a Human

The Boticário Approach: Personalization as Empathy

Grupo Boticário demonstrates a different form of humanization: an AI that shows empathy by deeply understanding the individual needs of each consumer.

Personalization as Empathy: - The AI remembers the customer's skin type and previous preferences - Recommendations consider occasions (gifts, daily use, special events) - The AI explains why it's recommending each product — educating rather than just selling - Adapts over time based on feedback and purchase history

This approach recognizes that one of the most valued aspects of human service is being understood and remembered. An in-store salesperson who remembers your skin type and last purchase creates a connection — and Boticário's AI replicates this digitally.

The +46% increase in conversion isn't just a matter of efficiency — it reflects consumers feeling that the experience is personalized and relevant, not generic and automated.

The lesson is: personalization is the digital form of empathy. AI that demonstrates it knows and understands the individual customer creates an emotional connection, even being non-human.

Want results like these for your business?

Talk to our team and see a platform demo

Satisfaction Metrics as a Compass

The Synthesia Approach: Self-Service with Smart Escalation

Synthesia demonstrates that even in a primarily self-service model, humanization is essential — not in the conversational AI itself, but in the intelligence of knowing when to activate human support.

Key Metrics: - 98.3% of queries self-resolved by AI - +33 percentage points in CSAT - 55% resolution rate

Paradoxically, CSAT increased 33 percentage points even with the vast majority of interactions being resolved by AI (not humans). This suggests that for many cases, quick and accurate self-service is more satisfying than waiting in a queue for human service.

The critical differentiator is the quality of escalation: the 1.7% of cases that reach human agents arrive with complete context, emotional analysis, and recommended solutions. This means the human agent can focus on what they do best — demonstrating empathy and critical judgment — without wasting time on triage and information gathering.

The lesson is: humanization isn't always about more human contact — it's about the right human contact at the right time.

How Catalisa Addresses This Scenario

Catalisa enables the creation of humanized AI agents that embody these principles — with configurable personas, conversation memory, intelligent escalation rules, and continuous monitoring.

With Catalisa's AI Agents, it's possible to design agents with distinct personalities for each brand or audience segment, configure intelligent escalation rules based on emotional and contextual signals, and monitor CSAT and satisfaction in real time.

The platform combines automation efficiency with customizable humanization, allowing companies to find the ideal balance for their specific context and audience.

See the platform

Bibliographic References

IT FORUM. “Carrefour aposta em IA para atendimento humanizado”. itforum.com.br. Disponível em: https://itforum.com.br/noticias/carrefour-aposta-ia-atendimento-humanizado/. Acesso em: 27 fev. 2026.

MUNDO DO MARKETING. “Grupo Boticário lança assistente de IA e eleva conversão em 46% no e-commerce”. mundodomarketing.com.br. Disponível em: https://mundodomarketing.com.br/grupo-boticario-lanca-assistente-de-ia-e-eleva-conversao-em-46-no-e-commerce. Acesso em: 27 fev. 2026.

FIN.AI. “Synthesia customer story”. fin.ai. Disponível em: https://fin.ai/customers/synthesia. Acesso em: 27 fev. 2026.

Want to replicate results like these?

Schedule a demo and see how Catalisa can transform your business

Related Articles

© 2026 Catalisa. Todos os direitos reservados.

Conforme LGPD

Meta WhatsApp Business

Dados no Brasil

We use cookies for analytics and load external resources (images) to improve your experience. By clicking "Accept", you consent to the use of cookies and third-party resources in accordance with our Privacy Policy.