Every Call Carries Brand Weight 

The world’s largest cosmetics corporation runs a CX operation where every interaction carries brand weight. Customers calling in carry high expectations. A premium brand lives or dies on how agents show up in every conversation, and soft skills are what determine that.

Traditional training had a ceiling:

• Classroom instruction built knowledge, not conversational instinct
• Human-led roleplay was limited by facilitator availability and session consistency
• Agents arrived at live calls technically prepared but underdeveloped in empathy and adaptability
• Feedback came too late to change behavior before the next real call

Practice Before the First Real Call

IntouchCX deployed Akin, its AI roleplay tool, to build conversational skills before agents ever touched a live call. Three things made it work differently than traditional methods:

• Realistic simulations. Agents practiced voice conversations with AI-generated personas built around the program’s actual top call drivers, not generic scenarios.
• Real-time scoring. Feedback on empathy, clarity, and resolution quality arrived immediately after each exercise, while the conversation was still fresh enough to act on.
• Curriculum integration. Akin was embedded directly into the existing new hire training flow. No separate module. No disruption to instructional design.

What the Numbers Showed

Measured against a control group on the same program. The AI-trained cohort outperformed the control group in every dimension.

ADOPTION: +49%usage rate, reaching 62% of agents

• AHT: 42 seconds faster than the non-Akin cohort
• CSAT: +6% improvement, more than 3x the control group’s gain

Adoption

• 954 production records logged
• 664 production + 1,265 training records in the final four weeks alone
• Voluntary usage. The adoption signal that matters.

Efficiency

• Entered with a wider AHT gap than the control group. Closed it anyway.
• Finished at 525 seconds AHT, competitive with the standard group.

Quality

• Peak CSAT of +89%, the highest the program has reached.
• Finished at +87%, on par with the non-Akin group.

What This Tells CX Leaders

Speed to proficiency is not a knowledge problem. Agents can know the product and still underperform on calls because the conversational skills were never built before they went live. What this program proved:

• Practice under realistic pressure develops skills that instruction alone cannot transfer
• Feedback speed determines whether a learning moment sticks or dissolves
• Akin-powered training does not replace human coaching; it multiplies the reps that make coaching land

For CX organizations running brand-sensitive programs, the training environment has to simulate what production actually demands. A model built for the easy version of the call will always produce agents who struggle with the hard ones.

Learn how to build conversational skills before day one with IntouchCX