← Cases
TaiwanFinance and insurance2018 – ongoing

Mega Bank

Xiao Mie AI customer service

Mega Bank reported that Xiao Mie handled 42% of customer-service traffic

WorkingCompany's own claim

Problem

Customers faced waits for telephone support, while the bank faced service staffing shortages.

Approach

Mega Bank launched Xiao Mie in 2018. It combined FAQ and natural-language-understanding models with APIs connected to internal databases. The service offered relevant follow-up options, such as credit-limit adjustments after available-credit inquiries.

Results

42%share of customer-service traffic handled by the assistant, up from 12% initially, according to Mega BankCompany's own claim[1]
96%response hit rate reported by Mega Bank; the bank did not define the measureCompany's own claim[1]

Takeaway

Database connections let the assistant support account-related inquiries and offer a relevant next action.

Categories

Business function
Customer service
Innovation type
Process innovation
Evidence
Company's own claim

Sources

  1. AI浪潮席捲金融業 兆豐智能客服小咩4成通話量靠她搞定
    Central News Agency, Howlife31 Jul 2023News article

Source summaries

[1]Central News Agency, Howlife31 Jul 2023

Mega Bank described Xiao Mie's models and database connections. The bank reported a 42% traffic share and a 96% response hit rate.

Full details are in the original sources.

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