ServiceNow Can Score Your Knowledge Articles. What Helps Authors Fix Them?

You finally have a number for your knowledge base. Maybe it's an AQI score from last quarter's review, or a list of findings from a Knowledge Center scan. Either way, it tells you something is off and roughly where. Then you open the first article on the list, and you're back to doing what you've always done: reading it top to bottom, guessing what "good" looks like for this team, and fixing it by hand. Here's how the review tools in ServiceNow and Amaze fit together, and where each one stops.

What Does ServiceNow Already Check in Your Articles?

More than most teams realize. There are two main tools, and each one does a different job.

• Article Quality Index (AQI): Your team writes its own checklist of weighted questions and assigns it to a knowledge base, so HR and IT can each be reviewed against their own standards (ServiceNow docs). A reviewer works through the checklist on a published article, and the true answers add up to a score (ServiceNow docs). It's the classic KCS review, run by knowledge coaches and domain experts.

• Article Optimization: Knowledge Center scans articles on a schedule. Script-based scans catch broken links, missing image alt text and heading problems. AI-based scans, which need Now Assist, judge things like whether a title fits the article. Findings show up in the advanced editor, and teams can write custom scans for brand, regulatory and writing standards (ServiceNow Community).

Together, these answer the questions a knowledge manager gets asked: how healthy is our content, and where do we start?

The condition: AQI reviews only reach the articles someone has time to review, and Article Optimization's AI-based and custom AI scans need Now Assist.

Where Does Review Still Get Stuck?

At the fix. Finding the problem is getting easier, while fixing it is still the same manual work it always was.

These are the patterns we see in onboarding:

• The finding arrives after the article exists. A scan runs on a schedule, so the author has usually moved on by the time it flags something.

• The rule doesn't know your team. HR wants plain language and a policy owner on every page. IT wants numbered steps and an error message in the title. One checklist rarely captures both.

• Correct isn't the same as useful. An article can pass every scan and still be a wall of text that nobody finishes reading.

• Reviewers become the bottleneck. When every fix routes back through a coach, the queue grows faster than anyone can clear it.

So the score can improve slowly even when everyone is working hard. If your articles are short and your standards are simple, scans plus a periodic AQI review may be all you need.

How Does AI Document Analysis Help Authors Fix the Article?

It brings the review to the author while they're still in the article, then helps them act on it.

Amaze AI Document Analysis (ADA) reviews the whole article on the canvas. When an author saves, Amaze offers to analyze it, and authors can run it any time from the AI agent. ADA returns a short summary and a few suggested improvements, measured against reference points tailored to your organization's standards, like the right length, a clear focus and the keywords your readers search for. The suggestions land in the AI agent, so the author can apply them with one prompt, such as "implement these changes, and don't change any styles." The author reviews the result, keeps what fits and undoes what doesn't.

That's the difference in where it lives. ServiceNow's scans measure the library over time. ADA works on one article at the moment someone is writing it, so problems get fixed before they ever become a finding. When you need to re-check a whole group of articles against new standards, the Amaze bulk tool runs that audit, and nothing republishes until someone approves it.

ADA runs on the AI provider your organization already approved, including Now Assist. Once that provider is connected, turning on analysis is one checkbox in Amaze Properties. Suggestions are a starting point, so a person still decides what's accurate.

Book a Demo

How Should You Use Both Together?

Let ServiceNow tell you which articles need work, and let Amaze help authors fix them.

A simple loop works well. Start with your scan findings and lowest AQI scores, focus on the articles your readers actually open, and hand those to their owners. Each owner opens the article in Amaze, runs the analysis, applies what fits, and republishes. The next scan or review shows whether it worked.

Write your standards once, then use them in both places. If your HR knowledge base's AQI checklist asks for a policy owner on every article, the reference points behind ADA should ask for the same thing. That way, authors and scans aren't working toward two different definitions of good.

One caution: a better score doesn't prove an article helps anyone. Keep an eye on article feedback and ticket volume, too.

Here's a ten-minute start. Pick your three most-viewed articles. For each one, write down the two things your team's standards say every article must have, then check whether it does. That short list is the first draft of the standards your analysis should check against.

Frequently Asked Questions

Does ADA replace Article Optimization or AQI reviews?
No. ServiceNow's tools measure your library, and ADA helps authors improve one article while they're writing it.

Do we need Now Assist to use ADA?
No. ADA works with Now Assist or another AI provider your organization has approved.

Does ADA change the article on its own?
No. It suggests improvements, and the author decides whether to apply them and can undo any change.

Can different teams check against different standards?
Yes. The reference points ADA uses are tailored to your organization's standards, so they can reflect what each team expects.

How do we re-check articles that are already published?
Use the Amaze bulk tool to audit a set of articles against your updated standards. Every result waits for approval before it republishes.