IT Knowledge Management Analyst Job Description

An IT department solves the same problems again and again: a VPN client that drops after an update, a print queue that stalls after a driver push, a password reset that fails because the account is locked in two directories at once. An IT Knowledge Management Analyst makes sure each of those fixes is written down once, in a form the next person can find and trust, and that it stays correct after the next change. The analyst owns the knowledge base the service desk works from, along with the self-service portal that end users search before they open a ticket.

The work splits into three streams. The first is content. Analysts write and edit articles themselves, but much of their time goes to shaping what others submit: turning a second-line engineer's terse resolution note into numbered steps, adding the exact error text a user will paste into search, linking the right configuration item, and deciding whether a new article is needed or an existing one should absorb the fix. A known-error article for an open problem record reads differently from a how-to for a new laptop, and the analyst keeps a template for each.

The second stream is structure and governance. Someone has to decide which knowledge bases exist, who can read and edit each one, which articles need approval before publishing, how long an article lives before review, and who inherits it when the owner leaves. In ServiceNow that means knowledge bases, user criteria, workflows and valid-to dates. In Confluence linked to Jira Service Management it means spaces, page permissions, labels and templates. Analysts also watch the change calendar so that articles about a system get rewritten before the upgrade, not after the first wave of calls.

The third stream is measurement. Analysts pull the search terms that return nothing, the articles agents attach to tickets, the ones users rate as unhelpful, and the incident categories with no linked article at all. That data feeds the content backlog. Public job classifications describe the same loop. The County of Santa Clara's IT Knowledge Manager series identifies knowledge and skill gaps by analyzing sources that include service desk data, scheduled projects and technology roadmaps, and its associate-level class monitors KM performance and develops monthly KPI reports.

Teams that adopt KCS, a methodology published by the Consortium for Service Innovation, aim to build knowledge capture and reuse into the way work gets done, so agents record and improve articles while they resolve requests instead of leaving documentation to a separate team afterward. In that model the analyst acts as coach and editor: setting the content standard, reviewing agent contributions, and repairing the workflow when capture drops off.

The role sits close to the service desk without being a support job. An analyst needs enough grounding in Active Directory, endpoint management, networking or the business applications being documented to spot a missing step or a risky shortcut. The rest is editorial judgment and persistence: getting busy engineers to review their own articles, retiring content that people are attached to, and keeping a large library honest.

Education. A bachelor's degree is the usual starting point, typically in information systems, information or library science, communications, or an IT discipline. Public classification specs show how one employer writes the requirement. The County of Santa Clara's Associate IT Knowledge Manager class calls for a bachelor's degree in business management, information management, library science or a related field, or equivalent training and experience, plus two years of experience establishing and maintaining content or knowledge management systems. Treat that as a single county's standard, not a national rule.

Experience. As a rough guide, analysts arrive with two to five years on a service desk, in desktop support, or in technical writing for an IT audience. Service desk time teaches what an agent needs in the middle of a call: the error message, the shortest route to the fix, and the point at which to stop and escalate. Writers and librarians who move into the job bring taxonomy and editing discipline, then learn the incident, problem and change side of IT service management on the job.

Credentials worth naming.

• KCS v6 certification (Consortium for Service Innovation). The Consortium says current KCS v6 training and certification remains fully valid, and that going forward KCS certifications will be refreshed every two years.
• ITIL Foundation (PeopleCert). In ITIL (Version 5), PeopleCert organizes the 34 ITIL management practices into two groups: product and service management practices, and general management practices.
• ServiceNow Certified System Administrator (ServiceNow), for analysts who configure knowledge bases, workflows and user criteria on that platform rather than only writing in it.

Skills that separate strong candidates.

• Plain-language technical writing: numbered steps, exact error text, one task per article, and titles phrased the way users search.
• Platform administration: knowledge base setup, permissions, approval workflows, templates and article lifecycle settings.
• Reporting: building dashboards from search logs, article feedback and ticket links, then turning them into a ranked backlog.
• Interviewing: drawing tacit knowledge out of an engineer in a short conversation without making it feel like an audit.
• Content judgment for AI features: spotting duplicate, conflicting or stale articles that could mislead a generated answer, and deciding which to merge or retire.
• Change awareness: reading the change calendar and release notes to see which articles a release will break.

Preparing for the move. A sanitized portfolio is worth assembling before applying: two or three before-and-after articles, a template you designed, and a short note on how you would measure whether a knowledge base works. Candidates coming from the service desk can point to articles they wrote that colleagues reused; candidates coming from writing should show they understand ticket data and the difference between an incident and a problem.

BLS has no separate category for this job, so the nearest official projection is the one ONET publishes for the group that holds knowledge management titles. ONET files Knowledge Management Specialist under Document Management Specialists (15-1299.03), part of Computer Occupations, All Other, and rates that occupation's projected growth for 2025 to 2035 as faster than average, at 5% to 6%, with about 27,000 openings a year on average. Those figures cover the whole group, which also includes IT project managers and computer systems engineers, so they describe the neighborhood rather than this title.

Setting matters for pay inside that group. In the May 2025 BLS industry data for Computer Occupations, All Other, the median wage was $128,670 in the federal executive branch and $75,150 in local government, excluding schools and hospitals. BLS does not report knowledge management analysts separately within either row, so read those medians as context for public-sector offers, not as figures for this job.

The KCS methodology for service desk knowledge work also changed in 2026. The Consortium for Service Innovation announced in April 2026 that KCS, formerly Knowledge-Centered Service, has become Knowledge-Centered Success, describing the change as a move from viewing knowledge management as a function of support and service to a fundamental driver of success across the entire enterprise. The Consortium ties the change partly to AI, writing that many organizations are discovering that their content isn't ready, because inconsistent structure, outdated information and lack of trust slow people down and break AI experiences. Its new Knowledge-Centered Success Practices Guide, released in Q2 2026 on the Consortium's timeline, gives guidance on keeping knowledge findable, accurate and relevant in light of new AI and automation tools. That is an advocacy body's argument rather than a measurement, but it names the problem analysts spend their days on.

ServiceNow has built article drafting into its platform. ServiceNow's Brazil-release documentation, updated September 10, 2026, describes ServiceNow Otto in Knowledge Management, which uses artificial intelligence to generate a knowledge article based on the resolution provided for a case or an incident, with versions for IT Service Management, Customer Service Management, HR Service Delivery and Field Service Management. Where agents use a feature like that, the analyst's review work can shift from fixing rough human drafts to checking generated text against what actually fixed the problem, deciding which drafts deserve publication, and keeping near-duplicates out of the library.

None of that removes the need for someone to own the knowledge base. A generated article still needs an owner, a review date, a place in the category structure and a retirement rule, and a self-service answer is only as current as the articles behind it.

For an individual job search, the useful question is how an employer treats knowledge. Ask in interviews who owns the knowledge base, how articles get retired, which AI features read from it, and whether the analyst has admin rights on the platform. The answers show whether the role is a service desk side task or a funded function with authority over content.

Dear Hiring Manager,

Please consider me for the IT Knowledge Management Analyst opening on your IT service management team. For the past four years I have worked on the service desk at a regional health system, and for the last two of them I have owned our knowledge base in ServiceNow alongside my queue.

When I took over the knowledge base, agents had stopped using it. Search returned three versions of the same VPN article, two of them wrong, and nobody knew who owned what. I set up a simple rule: every article has a named owner, a review date and a template that matches its type. I merged duplicates, retired content for systems we had already decommissioned, and rewrote our top password and multifactor articles so the first line of each one quotes the exact error a user sees.

The bigger change was cultural. I started reviewing agent drafts within a day, with specific notes rather than rejections, and I sat in on our weekly change meeting so that I knew which articles a release would break before it shipped. Agents began linking articles to their tickets again because they trusted what they found.

I also built a monthly report from our search logs and article feedback that shows the questions people ask and do not get answered. That report now sets my backlog, and our service desk manager uses it in planning.

I hold KCS v6 certification and ITIL Foundation, and I have been testing our platform's AI article drafting in a sandbox to learn where it needs a human edit. I would welcome the chance to bring the same discipline to a larger knowledge program.

Sincerely, Jordan Ellis