customer_support · saas · workflow

Grammarly achieves 87% deflection and 4.2 CSAT with Forethought

Grammarly's previous chatbot could not understand conversational context, broke when users rephrased or asked follow-up questions, and required the support team to manually author every answer inside a decision tree that was hard to manage and update.

How it works
Common implementation structure
How this type of workflow is generally built, generalized across documented cases — not tied to any one vendor's stack. Click any stage to read what happens there. Specific products that implement these stages appear in “Tools commonly seen” below.
Stage 1 · Customer contacts via email or chat
Grammarly's support team uses email and chat to connect with customers who need help.
Tools used
ForethoughtSolve
Outcome

After deploying Forethought, Grammarly's CSAT tripled to 4.2 out of 5 and deflection climbed from around 60% to a sustained 87%, never dropping below 80%. API integrations drove an additional 5–10% gain.

What failed first

Grammarly's prior chatbot relied on menu-based interactions and canned responses, lacked multi-turn context awareness, and forced the team into constant manual maintenance of a sprawling decision tree.

Results
Time saveda week and a half
Volume4.2 out of 5
Source

https://forethought.ai/case-studies/grammarly-achieves-87-deflection-and-4-2-csat-early-with-forethought

How we source this →

Grounding & classification
Source type: vendor customer story
33 fields verified against source quotes.
agentic workflowai agentconversational aiknowledge searchemailknowledge basesupport ticketfailure mode describedhuman review describedmetric backednamed customerproduction runtime claimedsource backedtools describedworkflow describedsoftwareautomation ratecustomer satisfactiondeflection rateemployee productivityvendor customer storycustomer supportticket triageautonomous resolutionintake to triage