Customer support ·

Jerome's Furniture cuts cost per contact 40% and average speed to answer 97% with Kustomer

The problem

Jerome's Furniture's previous ticketing-based platform couldn't keep up with the post-pandemic surge in digital channels — SMS and social — while call volume fell dramatically. Agents had to jump between systems because the POS wasn't integrated, slowing service and frustrating staff across all non-store channels.

First attempt

Their previous ticketing system was rigid and channel-siloed, unable to handle multichannel post-pandemic volumes. Agents were forced to jump between disconnected systems, and there was no structured way to capture reverse-logistics operational data.

Workflow diagram · grounded in source

Multichannel customer contact

trigger

“She oversees the entire post-sale customer journey — from delivery and service to warranty and issue resolution — across all non-store channels including phone, email, SMS, and social”

Unified agent dashboard

integration

“Kustomer was integrated across phone, SMS, social media, and email channels. This eliminated the need for agents to jump between systems and reduced response times.”

Conversations converted to Tasks

routing

“Jerome's Furniture converted conversations to Tasks in Kustomer. This change alone moved over 8,000 interactions into a lower-cost, more streamlined workflow that freed agents from repetitive steps and cut per-contact expenses. Shifting …”

Technician form submission

trigger

“Shop technicians now use tablets to fill out short Kustomer Forms as they complete each job”

Operational reporting output

output

“The reporting capabilities with Kustomer Forms gives unprecedented visibility into technician productivity and workload patterns. With these insights, Jerome's has been able to adjust workflows, allocate resources more effectively, and e…”

Reported outcome

Jerome's reduced cost per contact by 40%, achieved a 97% decrease in average speed to answer (from 32 minutes to under a minute), cut duplicated customer inquiries from 42% to 8%, and increased agent throughput to 12-15 touches per hour.
NPS reached two times the industry average.

Reported metrics

Reduction in customer inquiries50%

Decrease in cost per contact40%

Decrease in average speed to answer97%

Weekly call volume reduction75%

Show all 9 reported metrics

reduction in customer inquiries50%

decrease in cost per contact40%

decrease in average speed to answer97%

weekly call volume reduction75%

conversations shifted to Tasksover 8,000

duplicated customer inquiries42% to 8%

average touch per hour6-7 per hour to 12-15 per hour

average speed to answer32 minutes to under a minute

NPS vs industry averagetwo times higher than the industry average

Reported stack

KustomerKustomer Forms

Source

https://www.kustomer.com/customers/jeromes-furniture

Read source ↗

Frequently asked questions

What did this team achieve with this AI workflow?

Jerome's reduced cost per contact by 40%, achieved a 97% decrease in average speed to answer (from 32 minutes to under a minute), cut duplicated customer inquiries from 42% to 8%, and increased agent throughput to 12-…

What tools did this team use?

Kustomer, Kustomer Forms.

What results were reported?

Reduction in customer inquiries: 50%; Decrease in cost per contact: 40%; Decrease in average speed to answer: 97%; Weekly call volume reduction: 75% (source-reported, not independently verified).

What failed first in this deployment?

Their previous ticketing system was rigid and channel-siloed, unable to handle multichannel post-pandemic volumes.

How is this customer support AI workflow structured?

Multichannel customer contact → Unified agent dashboard → Conversations converted to Tasks → Technician form submission → Operational reporting output.

Jerome's reduced cost per contact by 40%, achieved a 97% decrease in average speed to answer (from 32 minutes to under a minute), cut duplicated customer inquiries from 42% to 8%, and increased agent throughput to 12-15 touches per hour.NPS reached two times the industry average.