Customer support · Production

Forethought Assist reduces D2L ticket close time by 13.7% and increases cases per hour by 32%

The problem

D2L support agents wasted time toggling between multiple windows and tabs to find help articles and past tickets, while keyword-based search tools returned long lists of irrelevant results from unstructured data sources, causing agents to spend more time searching than answering tickets.

First attempt

D2L's previous search tools relied on keyword matching and produced irrelevant results, and the team could not find a partner capable of working with their unstructured, heterogeneous knowledge sources.

Workflow diagram · grounded in source

Agent opens new case

trigger

“When an agent opens a new case in their helpdesk, Forethought Assist automatically pulls relevant past cases, knowledge articles, and macros relevant to the new case”

AI retrieves relevant content

ai_action

“Forethought Assist automatically pulls relevant past cases, knowledge articles, and macros relevant to the new case for the agent to use as reference, leveraging the collective knowledge of the entire support team”

Agent adds past response

human_review

“Using Forethought Assist's "Add to Reply" button, agents can add a past response to a new response with a single click”

Help article added to response

output

“the agent can easily add the help article into the response as well”

Knowledge gap identification

feedback_loop

“Using Forethought's cutting edge machine learning, D2L was able to identify gaps in their knowledge center so they could provide the right resources to help agents tackle the most common support issues in less time”

Reported outcome

D2L reduced time to close tickets by 13.7%, increased cases answered per hour by 32%, and decreased First Response Time by 56%, while agents who used Forethought consistently were 3.5 times more likely to meet their weekly efficiency goals.

Reported metrics

Time to close tickets13.7%

Cases answered per hour32%

First Response Time56%

AI accuracyover 90%

Show all 5 reported metrics

time to close tickets13.7%

cases answered per hour32%

First Response Time56%

AI accuracyover 90%

agent likelihood to meet weekly efficiency goals3.5 times more likely

Reported stack

Forethought AssistNLU

Source

https://forethought.ai/case-studies/d2l

Read source ↗

Frequently asked questions

What did this team achieve with this AI workflow?

D2L reduced time to close tickets by 13.7%, increased cases answered per hour by 32%, and decreased First Response Time by 56%, while agents who used Forethought consistently were 3.5 times more likely to meet their w…

What tools did this team use?

Forethought Assist, NLU.

What results were reported?

Time to close tickets: 13.7%; Cases answered per hour: 32%; First Response Time: 56%; AI accuracy: over 90% (source-reported, not independently verified).

What failed first in this deployment?

How is this customer support AI workflow structured?

Agent opens new case → AI retrieves relevant content → Agent adds past response → Help article added to response → Knowledge gap identification.