Customer support · Production

MPB scales recommerce customer support with Fin AI Agent, resolving up to 10,000 conversations monthly

The problem

As MPB expanded into new and existing markets, it needed to handle unpredictable volume surges—reaching 40,000 conversations in peak months—without simply growing headcount. Existing rule-based automation required ongoing maintenance and could not handle the nuance or complexity of MPB's unique customer journeys.

First attempt

MPB's prior rule-based Intercom automation helped streamline common queries but required ongoing maintenance effort and could not handle the nuance or complexity of their unique customer journeys.

Workflow diagram · grounded in source
1
Customer contacts via live chat
trigger
“Fin was introduced as the first point of contact in live chat, handling every inbound conversation”
2
Fin Guidance rules applied
ai_action
“A big part of that training is coming through Fin Guidance – a set of configurable rules that determine how Fin handles conversations. For MPB, this level of control is essential. Their inventory is constantly changing, with thousands of…”
3
Fin resolves query autonomously
ai_action
“Fin became a reliable first line of support – resolving thousands of queries quickly and accurately, while giving the team more time to focus on the personalized, high-touch conversations that drive value at MPB”
4
Empathy-triggered human handoff
routing
“If a customer seemed frustrated or the message felt urgent, Fin was guided to show empathy and hand it off to a human”
5
Ongoing Fin training and performance management
feedback_loop
“a Customer Operations Manager to oversee Fin's setup, training, and ongoing performance”
Reported outcome

Fin AI Agent now resolves up to 10,000 conversations per month and has doubled its resolution rate from 25–30%, while MPB maintains an 83% Customer Experience Score and delivers multi-lingual support in English, German, French, Dutch, and Italian without hiring native speakers in each market.

Reported metrics
Fin monthly resolutions (peak)10,000
Fin resolution rate48%
CX Score maintained83%
Resolution rate prior to Fin25-30%
Show all 7 reported metrics
Fin monthly resolutions (peak)10,000
Fin resolution rate48%
CX Score maintained83%
Resolution rate prior to Fin25-30%
Peak monthly conversation volume40,000
Team pressuretaken a lot of pressure off the team
Scale without headcount growthscale without simply adding more people to the team
Reported stack
Fin AI AgentIntercomFin GuidanceLLMsCX Score
Source
https://www.intercom.com/customers/mpb
Read source ↗

Frequently asked questions

What did this team achieve with this AI workflow?

Fin AI Agent now resolves up to 10,000 conversations per month and has doubled its resolution rate from 25–30%, while MPB maintains an 83% Customer Experience Score and delivers multi-lingual support in English, Germa…

What tools did this team use?

Fin AI Agent, Intercom, Fin Guidance, LLMs, CX Score.

What results were reported?

Fin monthly resolutions (peak): 10,000; Fin resolution rate: 48%; CX Score maintained: 83%; Resolution rate prior to Fin: 25-30% (source-reported, not independently verified).

What failed first in this deployment?

MPB's prior rule-based Intercom automation helped streamline common queries but required ongoing maintenance effort and could not handle the nuance or complexity of their unique customer journeys.

How is this customer support AI workflow structured?

Customer contacts via live chat → Fin Guidance rules applied → Fin resolves query autonomously → Empathy-triggered human handoff → Ongoing Fin training and performance management.

Related customer support cases