AI for Arabic banking and microfinance operations.
ChatSheba answers questions about products and application requirements using the bank’s approved information and collects enquiry details. Eligibility and financing decisions remain with the responsible team.
The problem in banking operations
Customers ask about Murabaha, Ijara, and Musharaka financing, account requirements, and daily exchange rates — patiently and repeatedly, often outside branch hours. Phone lines and relationship managers absorb questions that have the same answer every time, and Sharia-aware language has to be exact.
From an eligibility question to a financing file
- Eligibility enquiries: the financing type a customer asked about is recorded while the question is answered.
- Document preparation: the customer learns what is missing before the branch visit rather than after it.
- The customer learns which document an account-opening request is still missing without calling the branch.
- The phone team keeps the calls that need judgement once the agent takes the daily exchange-rate and remittance questions.
- Financing decisions remain with the responsible team, with access to information collected from the customer.
What the AI handles
- Explains Islamic financing products — Murabaha, Ijara, Musharaka — with eligibility and required documents.
- Answers account-opening requirements and daily FX and remittance questions.
- Maintains Sharia-compliant language and an Islamic greeting under a defined persona.
- Captures financing leads — type and amount — and routes them to relationship managers.
- Escalates underwriting and complex financial decisions to a human, with full context.
How Sheba fits
ChatSheba runs the customer-facing layer across WhatsApp Business API and the web; custom engineering connects it to your core systems where deeper integration is required. Both run inside your data boundary.
- See ChatSheba
Customer AI, through ChatSheba
Staff receive whatever needs a decision, after ChatSheba has handled the financing questions and applications arriving on WhatsApp and the web.
- See how AI agents work
Agents for back-office steps
Using only the tools it is permitted, an agent can check that a file is complete before it goes to review.
- See Sheba Enterprise Platform
An internal AI environment: Sheba Enterprise Platform
Evaluations are agreed one organisation at a time. It fits a bank whose staff need to search policies and product rules within their own permissions.
- Explore custom AI engineering
A custom system around core banking
Sheba builds a system to the same standards as its products when the work needs a deeper link to core banking.
Behind a financing file
- ChatSheba, for customer conversations on WhatsApp and the web.
- An AI agent with human escalation, so financing decisions reach staff.
- A product and eligibility knowledge base the agent answers from.
- Custom engineering where the core banking system needs a deeper link.
Compliance and data boundaries
Designed to support regulated deployment planning: per-deployment data boundaries, on-prem or regulated-residency paths available by scope, and human escalation as a built-in control.
- PDPL-aware by design
- On-prem and regulated-residency paths
- Per-deployment data boundaries
- Human escalation as a control
Composite industry example
A Sharia-compliant microfinance bank runs an Arabic AI agent under a defined name and persona. It opens with an Islamic greeting, explains financing products and eligibility, provides daily FX rates, and captures financing-application leads with type and amount — then hands underwriting to a human relationship manager. Leads reach the team pre-enriched.
How this pattern is set up
- Channels
- WhatsApp and a web chat widget
- Systems it reads and writes
- the core banking system and the customer record
- Escalates to
- a compliance officer on defined triggers
Part of 30+ production deployments across 14+ industries.
FAQ