Nedbank

Nedbank, one of South Africa's largest retail banks, faced a scaling challenge as digital banking adoption accelerated across its customer base. Growing volumes of routine inquiries — balance checks, transaction lookups, product questions, and service requests — were flooding human-staffed live chat channels. The bank needed to expand digital servicing capacity to match demand without a proportional increase in contact center headcount. Without an automated containment layer, every incremental user meant a direct cost increase and longer wait times, degrading the customer experience precisely as digital became the primary service channel.

Nedbank deployed Kasisto's KAI conversational AI platform as a branded virtual banking assistant named 'Enbi', embedded directly within its digital banking channels. Enbi uses Conversational AI & Virtual Assistants technology to handle the full range of routine customer interactions, including account information retrieval, transaction history queries, product inquiries, and common service requests. The assistant is integrated with Nedbank's core banking systems, enabling real-time account data access during conversations. Complex or sensitive matters are escalated to human agents via a smooth handoff, ensuring containment doesn't come at the cost of service quality. Kasisto provided the underlying financial-domain AI, which Nedbank branded and configured for its specific customer journeys.

Nedbank achieved a 70% reduction in live chat volume handled by human agents, with Enbi containing the majority of incoming digital customer interactions autonomously. This containment rate allowed the bank to scale digital servicing capacity in line with user growth without a corresponding increase in staffing.

Key outcomes:

• 70% live chat volume reduction — the headline containment metric
• Significant human agent deflection, freeing staff for complex, high-value customer interactions
• Digital banking channel capacity scaled to meet growing demand without headcount increases
• Improved response times for routine inquiries through always-on automated handling

• Branding the assistant matters — naming the bot 'Enbi' gave it identity and set clear customer expectations, smoothing adoption.
• Core system integration is non-negotiable — real-time access to account data is what separates a useful banking assistant from a glorified FAQ bot.
• Containment rate is the primary ROI lever — 70% deflection at scale translates directly to avoided headcount costs.
• Human escalation paths preserve trust — a well-designed handoff to live agents ensures AI containment doesn't create service dead ends.
• Financial-domain AI reduces build time — using a specialist vendor like Kasisto accelerates deployment compared to building general-purpose NLP from scratch.