Customer support · Production

Swyftx achieves 48.5% increase in automated resolution rate with Intercom Fin AI Agent

The problem

As Swyftx's user base surged, the volume and complexity of customer inquiries overwhelmed traditional support systems, and previous-generation rule-based bots failed to handle complex questions, creating customer frustration.

First attempt

Previous-generation chatbots built on rigid, rule-based structures could not handle complex questions, and customers were frustrated because they did not know the parameters of these rules.

Workflow diagram · grounded in source
1
Customer query arrives
trigger
“Fin's growing ability to handle customer queries of varying complexity levels on its own”
2
Clarify vague queries
ai_action
“Refine vague customer queries like "It's not working" by asking clarifying questions”
3
Generate contextual response
ai_action
“Generate accurate, contextual responses”
4
Validate before delivery
validation
“Fin's inbuilt safety mechanisms, which validate responses before they're delivered and safeguard sensitive customer data”
5
Continuous feedback loop
feedback_loop
“Regular reviews of Fin's conversations allowed Swyftx to identify gaps in responses and improve them over time”
Reported outcome

Fin AI Agent achieved a 49% resolution rate (a 48.5% increase over the previous solution), a 91% answer rate, and saved Swyftx's support team over 40 hours per week, with sentiment towards Fin improving dramatically.

Reported metrics
Automated resolution rate increase48.5%
Fin AI Agent resolution rate49%
Fin AI Agent answer rate91%
Team time saved per week40+ hours per week
Show all 5 reported metrics
automated resolution rate increase48.5%
Fin AI Agent resolution rate49%
Fin AI Agent answer rate91%
team time saved per week40+ hours per week
ramp to thousands of resolved conversations10 weeks
Reported stack
FinIntercom
Source
https://www.intercom.com/customers/swyftx
Read source ↗

Frequently asked questions

What did this team achieve with this AI workflow?

Fin AI Agent achieved a 49% resolution rate (a 48.5% increase over the previous solution), a 91% answer rate, and saved Swyftx's support team over 40 hours per week, with sentiment towards Fin improving dramatically.

What tools did this team use?

Fin, Intercom.

What results were reported?

Automated resolution rate increase: 48.5%; Fin AI Agent resolution rate: 49%; Fin AI Agent answer rate: 91%; Team time saved per week: 40+ hours per week (source-reported, not independently verified).

What failed first in this deployment?

Previous-generation chatbots built on rigid, rule-based structures could not handle complex questions, and customers were frustrated because they did not know the parameters of these rules.

How is this customer support AI workflow structured?

Customer query arrives → Clarify vague queries → Generate contextual response → Validate before delivery → Continuous feedback loop.

Related customer support cases