OBLYTECH Logo
OBLYTECH Logo

ServiceNow / Practical guide

ServiceNow Knowledge Deflection That Supports Incident Resolution

ServiceNow knowledge deflection is useful when a person finds accurate guidance and can complete the next step without opening avoidable support work, or when the article helps an agent resolve the incident faster. A page view alone does not prove that either outcome happened.

For ITSM and knowledge owners reducing repeated incidents through useful, maintained support guidance..

ServiceNow knowledge deflectionServiceNow knowledge managementincident deflection

Guide map

Linking Knowledge Articles to Incident Categories

Define deflection as a useful user outcome

The reader finishes the task or reaches the right support path.

Write around the question before the article title

Start from the question a person types, not from the article title.

Connect knowledge to incident and request workflows

Place each article where the question arises, such as intake or agent work.

The practical framework

Improve ServiceNow knowledge deflection by matching real incident questions to clear articles, ownership, search language, feedback, and escalation.

Start with repeated questions and failure points in the incident process. Build guidance around the decision the reader needs to make, the steps they can safely perform, and the point where they should request help.

Define deflection as a useful user outcome

Separate three outcomes: a reader solved the issue, a reader understood that the article did not apply and reached the right support path, or an agent used the article during resolution. All three can make knowledge useful, but they require different evidence and review.

The ServiceNow knowledge-management route is the relevant internal destination when article ownership, lifecycle, search, and review need a structured design.

Do not define success as hiding ticket creation. A person who cannot find the right support path may abandon the request or submit incomplete information. The goal is a clearer next action and a better resolution path.

Write around the question before the article title

Review incident descriptions, request wording, agent notes, and search terms that people actually use. Group questions by task, symptom, decision, audience, and risk. Then write the shortest useful answer that explains the action, prerequisites, expected result, and escalation point.

Use plain headings and terms that match the reader’s language. Explain platform terms where they are necessary. Keep troubleshooting steps ordered, reversible where possible, and explicit about when the reader should stop.

The ServiceNow employee knowledge route provides an internal path when knowledge must be connected to a wider employee or service experience.

Connect knowledge to incident and request workflows

Place relevant guidance where the question occurs: during search, incident intake, agent work, request fulfilment, or closure. The link or recommendation should match the record context. A generic article list is less useful than a small set of relevant next actions.

The ServiceNow incident-management route is the appropriate internal destination when knowledge should support triage, handoff, resolution, and closure.

Make escalation clear. State what information the person should include, what the support team will need to know, and which symptoms should not be handled through self-service.

Give every article an owner and review path

An article needs an accountable subject owner, a review trigger, a replacement or retirement decision, and a way to capture feedback. Review it when the supported workflow, access, system behaviour, policy, or assignment route changes.

Feedback should lead to a decision. A comment may show missing context, an unsafe step, a broken link, a changed screen, or a question that deserves a separate article. Record the disposition rather than collecting comments without action.

Measure usefulness without chasing a number

Review search terms with no useful result, repeated incidents, article feedback, escalations after article views, agent use, and questions that remain unresolved. Interpret each signal with context. A high article view count may mean the page is useful or that people cannot find a clearer answer.

The ServiceNow continuous-improvement route gives the reader an internal path for turning those findings into a recurring content and process review.

Common questions

Questions teams ask about knowledge deflection

Is every self-service view a deflected incident?

No. A view shows that the article was opened. Deflection needs stronger evidence that the reader completed the task, avoided avoidable support work, or reached a better support path. Use several signals and interpret them against the service context.

What should happen when an article does not solve the issue?

The reader should have a clear escalation route and know what information to provide. The support team should be able to record that the article was tried, what happened, and whether the article needs correction.

The research-grounded boundary

In ServiceNow, knowledge sits on the same platform as incidents, requests, and permissions, so a deflection decision depends on that context. Before changing an article or a deflection path, record what the reader may do, what they must still confirm with a person, and who owns the outcome. The decision should be traceable from its starting condition to its owner, its exceptions, the evidence behind it, and the date it will be reviewed.

Start from the failed search, not the article list

A very useful input for a knowledge program is not the catalog of existing articles. It is the set of searches that returned nothing useful, the incidents that were resolved by the same three steps every time, and the agent notes that begin with the same explanation. Those signals show where a reader is stuck and what words they use to describe the problem. An article written from the failed search has a far better chance of being found than one written from the team’s own vocabulary, because the title and opening lines already match how the reader thinks about the issue. Incident-side handling of the same symptoms is covered in incident and problem management.

Match the article shape to the question type

Not every question needs the same kind of article. A question about a symptom needs a diagnostic path that ends in either a fix or a clear escalation. A question about a request needs a short description of the steps, the approval involved and the expected wait. A question about policy needs a decision guide that says which case applies. Forcing all three into a single long how-to produces articles that are hard to scan and easy to misapply.

Question typeArticle shape that fitsWhere it should appearSignal it should produce
Symptom: something is not workingDiagnostic path with a stop point and an escalation routeSearch results and incident intakeFewer duplicate incidents for the same symptom
Request: how do I get somethingShort sequence with approvals and expected next stepCatalog item and request confirmationFewer requests returned for missing information
Policy: is this allowedDecision guide with the cases listed in orderLinked from the policy and from the related requestFewer exceptions raised for cases the policy already covers

A publishing checklist for each article

A checklist is more useful here than a style guide, because many weak articles fail on basic points that are easy to miss when the writer knows the subject well.

  1. Confirm the question the article answers, in the words a search would use, and check it against two or three recent failed searches.
  2. List the prerequisites the reader needs, such as access, a device state or a role, and state what to do when one is missing.
  3. Write the steps in order, mark any step that cannot be undone, and state the expected result after each risky step.
  4. Add the stop point: the condition under which the reader should stop and raise a request or an incident.
  5. Name the subject owner and the review trigger, such as a change to the supported workflow or the access model.
  6. Test the article with someone who did not write it, asking them to follow it on a real case and report where they paused.

Patterns that make deflection numbers look better than they are

A rise in article views can come from people who could not find the right support path and kept searching. A fall in incidents can come from a change in how the intake form routes requests rather than from the article. A high rate of article feedback that says helpful may mean the article is short rather than correct. None of these signals is wrong on its own, but each needs a second signal beside it, such as whether the same reader later raised an incident about the same symptom. A knowledge owner who reviews only one number will eventually optimize for it. For the wider ITSM picture, see the ITSM transformation guide.

What to review after the first period of use

Review search terms that returned no article, articles that were opened and then followed by an incident on the same topic, and feedback that reports a broken step rather than a general opinion. Each finding should end in a disposition: correct the article, write a new one for a question that has no home, retire an article that duplicates another, or change the intake route. Feedback that is collected but never dispositioned teaches readers that their input does nothing. For the service-operations context these signals feed into, see ITSM service operations.

Questions about knowledge deflection

Should every knowledge article be visible to all users by default?

Not necessarily. Articles that describe privileged steps or internal escalation contacts should carry an audience restriction, while the symptom and request articles usually belong to a wider group. Decide the audience from the risk of a reader acting on the wrong instruction.

How should an article that applies to one region only be handled?

State the region, the system or the business unit in the title or the first paragraph. A regional variant that is hidden inside a general article causes readers in other regions to follow steps that do not apply to them.

Can an agent edit a knowledge article during an incident?

An agent can propose a correction during an incident, but the change should go through the subject owner’s review, because a fast edit made under pressure is a common source of an incorrect instruction staying live.

What belongs in the feedback form?

Ask whether the article solved the issue, which step was unclear or wrong, and whether the reader still needs help. Avoid asking for a rating alone, since a rating does not tell the owner what to change.

Reviewing the questions your team handles most

If your knowledge base has articles that are opened but do not stop incidents, a review of the search terms and incident links can show where the gap sits. The knowledge and self-service optimization page covers connecting real questions to clear answers, or you can get in touch about the questions your team handles most often.

A practical next step

Need a clearer ServiceNow operating model?

OBLYTECH can discuss the questions raised by this guide and the next review step. See the ServiceNow consulting services page for the commercial route.

Useful starting pointBring the current question, scope, ownership concerns, and evidence needs into the first review conversation.

ServiceNow guides on this topic

Guides from the Oblytech ServiceNow library that explain the practice behind this page.

More from the blog