Customer support · Production

Forethought Solve and Triage help Achievers reach 69% deflection rate and 93% first contact resolution

The problem

Achievers' support queue was backed up with repetitive, low-complexity inquiries — 22% were password resets — assigned to agents without context, while a homegrown chatbot provided no deflection capability and no meaningful ticket classification.

First attempt

Achievers' previous homegrown chatbot lacked the ability to deflect simple inquiries, causing the agent queue to back up with tickets that should have been self-served.

Workflow diagram · grounded in source
1
Support inquiry arrives
trigger
“Achievers experiences a spike in support inquiries around the busy holiday season”
2
Triage predicts and classifies ticket
ai_action
“Triage uses historical data to proactively predict and classify new incoming tickets for Achievers. Achievers uses Triage to predict the reason the support ticket was submitted, across over 100 languages.”
3
Spam detection deflects spam
routing
“The spam detection feature automatically deflects and eliminates spam tickets that would otherwise clog up the support queue”
4
Solve resolves repetitive tickets
ai_action
“Solve instantly searches Achievers' entire database of knowledge articles and previously resolved tickets to provide the most accurate responses to support inquiries, without the intervention of an agent”
5
Gap detection content feedback
feedback_loop
“With gap detection, Achievers is able to identify areas where content needs to be created and whether the current articles aren't working to answer questions”
6
CSAT metrics collected
output
“Achievers also collects CSAT metrics directly within the widget through a pop up that appears to customers after closing the chat window”
Reported outcome

Achievers achieved a 69% deflection rate with Solve (far exceeding the initial expectation of 10%), 93% first contact resolution, a 50% increase in engagement score, and eliminated the need for 5 support agent headcounts through natural attrition.

Reported metrics
Solve deflection rate69%
Triage and Solve combined deflection rate44%
First contact resolution rate93%
Engagement score increase50%
Show all 9 reported metrics
Solve deflection rate69%
Triage and Solve combined deflection rate44%
First contact resolution rate93%
Engagement score increase50%
Support agent headcounts eliminated5
Password reset inquiries share22%
Initial deflection rate expectation10%
Best-case deflection rate expectation30%
Languages supported by Triageover 100 languages
Reported stack
ForethoughtSolveTriageDiscoverSalesforce Service Cloud
Source
https://forethought.ai/case-studies/achievers
Read source ↗

Frequently asked questions

What did this team achieve with this AI workflow?

Achievers achieved a 69% deflection rate with Solve (far exceeding the initial expectation of 10%), 93% first contact resolution, a 50% increase in engagement score, and eliminated the need for 5 support agent headcou…

What tools did this team use?

Forethought, Solve, Triage, Discover, Salesforce Service Cloud.

What results were reported?

Solve deflection rate: 69%; Triage and Solve combined deflection rate: 44%; First contact resolution rate: 93%; Engagement score increase: 50% (source-reported, not independently verified).

What failed first in this deployment?

Achievers' previous homegrown chatbot lacked the ability to deflect simple inquiries, causing the agent queue to back up with tickets that should have been self-served.

How is this customer support AI workflow structured?

Support inquiry arrives → Triage predicts and classifies ticket → Spam detection deflects spam → Solve resolves repetitive tickets → Gap detection content feedback → CSAT metrics collected.