Bank Hapoalim
Overview
Section titled “Overview”| Field | Detail |
|---|---|
| Industry | Banking |
| Country | Israel |
| Channel | Voice + chat |
| Initial delivery | 72 hours for earlier appointment agent; 3 weeks for later savings-campaign agent |
| Primary use case | Appointment scheduling and an eligibility-driven savings campaign |
Business challenge
Section titled “Business challenge”The bank wanted more effective customer self-service, but also needed its own staff to build and update agents without permanent vendor dependency. Banking workflows required identity, eligibility, and messaging integrations.
AI solution
Section titled “AI solution”The earlier appointment agent handled scheduling and validation. A later savings-campaign agent explained an offer, verified identity using ID/OTP, checked eligibility, confirmed participation, sent confirmations, and escalated when required.
Technical implementation & delivery
Section titled “Technical implementation & delivery”The later agent used a RAG knowledge source and 15 custom tools linked to bank systems. Wonderful’s forward-deployed engineers built reusable integrations that an in-house data engineer adapted for the second agent.
Results & KPIs
Section titled “Results & KPIs”| Metric | Reported result | Measurement scope / context |
|---|---|---|
| Earlier appointment-agent launch | 72 hours | Initial use case only |
| Earlier agent resolution | 75% | Appointment agent; separate deployment |
| Earlier customer sentiment | 97% positive | Earlier appointment-agent case |
| Later campaign-agent build | 3 weeks | Bank-owned development using platform and tools |
| Later voice containment | 88% | Savings-campaign agent |
| Campaign audience | ~100,000 eligible customers | Target segment, not achieved conversions |
Key takeaway
Section titled “Key takeaway”The differentiation is FDE-built reusable tools plus transfer to the customer’s own engineering team.