Response time down to under two minutes
Esportes da Sorte blends AI and human support to scale empathy
Esportes da Sorte reworked its customer service from the ground up-pairing AI assistance with trained agents-to speed up help without losing the human touch. Head of Customer Service Maria Neves summed it up clearly: "Artificial intelligence wasn't designed to replace people, but to empower them."
The results are hard to ignore. After restructuring, rolling out a new CRM, and retraining staff, average response time dropped from 30 minutes to under two minutes. As Neves put it, "We achieved operational efficiency without sacrificing empathy."
What changed under the hood
- AI-assisted triage and guidance: Predictive routing and suggested responses help agents move faster while keeping ownership of the interaction.
- New CRM + streamlined workflows: A unified view consolidates context, so agents don't waste time clicking through tools.
- Targeted training: Teams learned how and when to use AI, how to personalize, and how to spot responsible gaming signals.
- Clear scorecard: NPS, CESAT, and resolution time drive daily coaching and continuous improvement.
Results that matter to support teams
- Speed: Average response time now under two minutes.
- Quality: Faster replies without losing empathy and context.
- Trust: Each interaction is treated as a chance to reinforce responsible use and long-term loyalty.
Responsible by design
The operation follows Brazil's Law No. 14.790/23 and guidance from the Secretariat of Prizes and Betting (SPA/MF). The company partners with Instituto de Apoio ao Apostador (IAA) and Empresa Brasileira de Apoio ao Compulsivo (EBAC) to reinforce responsible gaming, and it holds Great Place to Work certification-signaling a culture that supports both customers and agents.
All of this adds up to a simple strategy: use tech to anticipate needs and personalize, keep people at the center, and make compliance and data governance non-negotiable. That's how you build something transparent and sustainable.
What you can apply in your own team
- List your top 10 contact intents. Automate the simple ones with guardrails and route edge cases to skilled agents.
- Set confidence thresholds for AI. If confidence is low or risk is high, hand off to a human immediately.
- Unify context in your CRM. Enforce tags, notes, and templates so every agent sees the same picture.
- Track NPS, CESAT, first response time, and full resolution time by queue and time of day. Review daily; coach weekly.
- Train agents on prompt patterns, tone, and judgment. Run side-by-side tests before rolling changes to production.
- Bake in responsible-use checks and data policies. Log decisions, permission access, and audit regularly.
If you're upskilling your support org on practical AI, browse curated options by job role here: Complete AI Training.
Your membership also unlocks:
