Article Engagement
Per Article: ViewCount, unique-Customer reads, HelpfulRatio, average DwellSeconds, feedback-comment volume. Drills Category × ContentType × Owner. Pairs with the Stale-Article queue as the KB Manager's daily landing report.
Helpdesk and insights
Reuse support answers and measure what happened after customers read them.
OutcomePendingOutcomeNoTicketObservedTicketCreatedAnywayUnknown
| Event ID | Article Version | Observation Start | Observation End | Connector Coverage | Outcome |
|---|---|---|---|---|---|
| EI-340 | AV-9325 | 12 Sep | 17 Sep | Complete | PendingOutcome |
| EI-494 | AV-2626 | 23 Sep | 02 Oct | Incomplete | NoTicketObserved |
| EI-351 | AV-1923 | 19 Sep | 28 Sep | Unknown | TicketCreatedAnyway |
| EI-387 | AV-5711 | 20 Sep | 22 Sep | Complete | Unknown |
| EI-247 | AV-1958 | 10 Sep | 17 Sep | Incomplete | PendingOutcome |
| EI-807 | AV-2824 | 15 Sep | 22 Sep | Unknown | NoTicketObserved |
Promote an answer without publishing a ticket
A resolved Helpdesk interaction can contain knowledge worth reusing. A trusted source event can propose one draft with its external ticket identity and a stable event key. The Helpdesk remains the ticket owner; Knowledge Base Platform does not create a competing ticket table or rewrite the resolution. Repeated promotion events return the original candidate.
The draft still needs an authorized author. A support agent who can resolve tickets does not automatically receive article-authoring permission. Missing author mapping stays visible for KB Manager assignment. Customer-specific details must be removed before the candidate enters ordinary article review, and the eventual public answer should stand on its own without exposing the original customer case.
Separate agent use from customer self-service
The Helpdesk widget can record which published article helped an agent prepare a reply. That is a useful knowledge-use signal, but it is different from a customer reading an answer before creating a ticket. The application keeps those populations separate so a busy agent using the library does not artificially improve the self-service outcome rate.
For a permitted correlated customer session, the first eligible read anchors one observation window and the exact article version. Additional reads in that same window do not each create another success opportunity. Anonymous or unlinked sessions remain outside this correlated outcome measure, while their permitted aggregate usage can still inform article engagement.
Report an observation rather than a causal claim
A verified ticket-creation event within the configured window records that a ticket followed the read. A no-ticket-observed outcome requires the window to close and the connected Helpdesk event feed to confirm complete coverage through the declared lag. If that evidence is missing, the outcome is unknown. A connection outage cannot become a successful self-service result.
Reports show pending and unknown observations separately from completed windows. The resolved rate compares no-ticket-observed with ticket-created outcomes on that same basis; it does not prove that an article prevented a ticket or created financial savings. KB Manager can read this alongside search effectiveness, article helpfulness, review time and contribution history. Keeping each source and observation boundary visible gives the team useful evidence for improvement without inflating the result when customers browse several pages or an integration falls behind.
Modules
-
Article authoring
Turn a useful answer into a structured draft with the context readers need.
-
Review and publishing
Approve the exact article revision and publish a frozen version.
-
Search and self-service
Help customers and support agents find the current answer they can access.
-
Content gaps and feedback
Turn unanswered searches and reader feedback into accountable content work.
-
Knowledge maintenance
Keep review dates and broken-link work visible after publication.
-
Helpdesk and insights
Reuse support answers and measure what happened after customers read them.
Reports
All reportsDeflection Rate
Daily / weekly / monthly NoTicketObserved count divided by (NoTicketObserved + TicketCreatedAnyway). Pairs the KB read volume against the Helpdesk ticket-create volume from sessions that touched the KB. An observed session outcome, not a causal ticket-saving or ROI claim. Unknown/incomplete windows are excluded from the resolved rate and reported separately.
Author Contribution
Per Author: Articles Created, Edited, Reviewed, Approved, Published over a rolling 90-day window, plus average HelpfulRatio of own Published Articles. Drills from Author → Article. Powers the contributor leaderboard.
Roles and permissions
Drafts and maintains assigned knowledge articles.
Searches approved knowledge while helping customers.
Finds and rates current public answers.
Related processes
Draft to published answer
Move an exact article revision through the required reviews into publication.
4 stages · 1 approval
Search to useful answer
Deliver the permitted published answer and capture its actual reading context.
4 stages · 0 approvals
Content gap to reviewed update
Turn search gaps, ratings and overdue review into a published improvement.
4 stages · 1 approval