Finance / Banking
Agibank: Identity-verified WhatsApp loan status responses and escalation of risky inquiries
- Company
- Agibank
- Country
- Brazil
- Adoption stage
- In operation
- Source published
- Date basis
- The date the source was published. It can differ from the date adoption started.
- How the source was checked
- Read the full source text
The work problem
Agibank wanted to grow its customer base without increasing operating costs. More than 200 service representatives handled over 350,000 WhatsApp inquiries each month, many involving routine loan status questions. Making customers wait for a representative even for basic status checks added pressure to service expansion.
Technology and data
Agentforce greets customers in Portuguese and asks for their Brazilian identification number, CPF. A connected workflow verifies it against customer data in Service Cloud before routing to the loan status agent. The agent finds open opportunities and Order Object information, explains the status conversationally, and saves an interaction summary in the customer profile. Compliance or fraud risks trigger escalation to a representative. Staff use Tableau Next to monitor intent, resolution, routing accuracy and handoffs.
Results
Salesforce reports that Agibank's loan status agent autonomously handles 22,500 inquiries per month with a 75% deflection rate. Routine checks are automated while risky conversations or those needing other expert judgment go to people. The bank says this supports growing demand without adding headcount.
Limits and open questions
The 75% metric belongs to the loan status agent, not autonomous resolution of all 350,000 monthly inquiries. The source does not provide accuracy, changes in customer satisfaction or quantified operating cost savings.
Sources
- Agibank answers 22,500 loan status inquiries per month with AgentforceSalesforce, Accessed
Compiled from public sources. These are not results from ATF Works customers.