← Cases
BrazilFinance and insurance2023 – ongoing

Sicredi

Generative AI assistant for consortium-product support

Sicredi's support pilot increased query resolution without specialist escalation by 10%–12%, IBM reported

WorkingVendor case studyVendor ↗

Problem

Support representatives often needed product specialists to resolve member questions. Escalations increased waiting times.

Approach

Sicredi and IBM built an assistant using watsonx.ai, Watson Discovery and watsonx Assistant. It generated answers from support documentation. A 20-day pilot covered consortium-product enquiries and was compared with the previous month's support data.

Results

10%–12%improvement in queries resolved without a product specialist during the pilot, per IBMVendor case study[1]
8%decrease in support calls abandoned because of waiting time, per IBMVendor case study[1]

Takeaway

The reported improvement covered one product and a short pilot; expansion to other products remained a next step.

Categories

Business function
Customer service
Innovation type
Process innovation
AI technology
Generative AI
Evidence
Vendor case study

Sources

Source summaries

[1]IBMDec 2023

IBM described an assistant tested on consortium-product enquiries. It reported better resolution without escalation and fewer abandoned calls against the preceding month's results.

Full details are in the original sources.

Added 1 Nov 2026 · Updated 1 Nov 2026
Open case page ↗