Back to cases
WillbankCustomer service & contact center · Digital bank

Wiki and AI for faster, more accurate service

Willbank centralized its processes in the Beedoo Wiki and put Abigawill, its AI assistant, to answer the operation's questions. The results show in the service metrics.

AHT in May
25:09

Average handling time before the Wiki and Abigawill were consolidated.

AHT in July
22:21

About 3 minutes faster average handling time.

The solution

Knowledge in the hands of agents

Process Wiki

Centralized procedures, viewed more than 300K times a month by the operation.

Abigawill, Willbank's AI

An assistant that answers agents' searches based on Wiki content.

Operation announcements

Announcements tracked month by month, with an 80% readership target.

Tom Santos — Willbank
Customer service & contact center · Digital bank
Willbank
“Using the Wiki and the Abigawill ChatBot for efficient customer service.”
Tom Santos, Willbank
Results

What changed in the metrics

−3 min
in average handling time
77.1%
correct answers from Abigawill
0.61%
of contacts became tickets (was 1.66%)
95%
announcement readership (target 80%)
Case screens

The numbers, month by month

Willbank TPR, AHT and TR chart for May and July

Service metrics

Between May and July, AHT dropped from 25:09 to 22:21, resolution time from 25:35 to 22:35 and response time from 26 to 21 seconds.

Abigawill results chart

Abigawill overall results

Across 8,362 searches, Abigawill was right in 77.1% of cases, combining direct answers (50.5%) and disambiguation (35.6%). Only 22.9% were not understood.

Willbank chats and tickets chart, May to July

Chats and tickets

With more chats (57,833 in July), tickets fell from 853 to 352 — from 1.66% to 0.61% of contacts.

Announcement readership chart, January to August

Operation announcement readership

Readership went from 56.8% in January to above the 80% target in July (95%) and August (92%).

Next step

Want this result in your operation?

In 30 minutes we show how Wiki, AI and communication work together in your service context.