nav[aria-label="Primary Navigation"] { padding: 0; & ul { list-style: none; width: 100%; display: flex; flex-direction: row; justify-content: start; align-items: start; gap: 30px; padding: 0; & li { margin: 0; } & ul li { list-style: none; } } }

Duplicates, conflicts, ownerless pages: how do you keep Confluence docs consistent?

Artiom Aleksandrovici

Have you ever found two pages with different instructions and had to figure out which one was actually correct?

As documentation grows, duplicates, outdated information and conflicting instructions get harder to track. Some pages end up with no owner at all because the author left the company or lost access.

How do you deal with this on your team?

Do you run regular documentation audits, or fix things only when someone complains?
Who becomes the owner of a page when the original author leaves?
How do you find conflicting instructions across spaces?
Do you rely on built-in tools (automation rules, archiving, labels) or Marketplace apps?

Comments

Joshua Brock _ Seibert Group_ GmbH

Hey @Artiom Aleksandrovici, greetings and welcome!

I think there are a few different problems wrapped together here, and I wouldn't necessarily try to solve all of them with the same tool.

For pages whose original owner has left, Confluence already gives you some useful options. In Premium and Enterprise, Content Manager can identify pages whose owners are no longer active, and space admins can reassign ownership in bulk. That's probably where I'd start with the ownerless-page problem.

Duplicates and conflicting instructions are harder. Tooling can help you find stale or poorly maintained content, but ultimately someone still has to decide which page is authoritative and whether the other one should be updated, archived, or removed.

Where I'd look at adding a more structured process is to prevent that problem from building up again.

Full disclosure: I work for Seibert, the company behind Aura Workflow & Approval.

Aura Workflow & Approval can be used to create recurring content review cycles. For example, a page could remain in an Up to Date state for a defined period and then automatically move to Needs Review. The appropriate reviewer is notified and can confirm that the content is still accurate, update it, or move it toward an outdated/archive state.

That changes content maintenance from “we'll audit everything when somebody remembers” into an ongoing lifecycle.

You can also use workflow actions such as labels, notifications and approvals, and reviewers have an Open Tasks view for content currently awaiting their attention. That makes it easier to see what actually needs review instead of relying entirely on periodic manual audits.

I wouldn't claim that a workflow solves conflicting instructions automatically. Someone still needs to make that judgment. But a regular review cycle makes it much less likely that an abandoned page will quietly sit next to current documentation for years without anyone noticing.

So my approach would probably be a combination: use Confluence's native ownership and Content Manager capabilities for cleanup and reassignment, then add a recurring review workflow if you need stronger governance around keeping the content current going forward.

Hope this helps, best of luck and again, welcome!

Joshua
Content Writer & US Representative
Aura Apps and Agile Hive (products of Seibert Group GmbH)

Artiom Aleksandrovici

Hi Joshua, thanks for the thorough reply. The split between ownership cleanup (Content Manager) and ongoing review cycles is useful.

On conflicting instructions, you said someone has to make the judgment, and I agree. What I’m curious about is the step before that: how does anyone find out there’s a conflict?

A typical example: one page says customers can request a refund within 30 days, another page in a different space says 14, and a third copies the 30. Each page has its own owner, and each could be confirmed as “Up to Date” by its own reviewer, because each reviewer only reads their own page. Nobody compares them, so the review cycle may confirm both.

A few questions:

Detection. In practice, how do teams catch cross-page conflicts like this? Do they surface only when a colleague or customer notices, or is there a process or tool that compares pages against each other?
Deciding. When the owners sit in different teams, who has the final say on which page is authoritative? And once decided, do you replace the copies with an Include Page macro, or edit each one?
Tools. Is there anything you’d recommend (on the Marketplace or elsewhere) that looks at what pages say across spaces rather than at dates and review states? Even a partial tool that flags likely conflicts for a human to review would be interesting.

Thanks again for the clear overview!

Murray le Roux

Hi @Artiom Aleksandrovici,

Your question describes the exact problem I am trying to solve with an app I built, so I hope this doesn't come across as too much self promotion but it feels relevant.

Comparing related pages is hard to do consistently as a knowledge base grows. Once a conflict is found, someone responsible for the underlying policy still needs to decide which guidance is correct.

I built Evergreen AI for Confluence. It reads your documentation and flags issues like this where pages disagree, information has become outdated, or ownership has become stale and lists it to you with the quoted evidence, allowing you to assign an owner for each issue it finds

It’s just launched on the Marketplace. If you give it a try and find something missing or a finding that’s wrong, I’d be happy to hear about it.

One limitation relevant to your example: comparisons currently happen within individual spaces, so it doesn’t yet address conflicts between pages in different spaces. If that's something that's very important to you, let's chat. I'd be happy to make adjustments.

Hope this helps!

Murray le Roux

Rezgar Cadro {Totem Dev}

Hi @Artiom Aleksandrovici ,

For specific problems like "make sure these pages are consistent" or "make sure no page contradicts on X" you can use Rovo. It's great at fixing a specific problem or specific pages, but not at full KB scale.

The problems you described are typical symptoms of a long-lived KB's health decaying. A healthy KB takes regular housekeeping across the whole KB, as it lives and breathes, and that's what we built WikiFix for.

WikiFix resolves each of the problems you described, and more. Your three refund pages become one question: which window is right? You answer it and all pages are updateed. Duplicates are merged into one page that keeps the important facts, and ownerless pages get a new owner. If your KB spans several Confluence instances or Notion workspaces, one scan covers all of them.

You choose the spaces to monitor. The first scan finds every problem already there, and a weekly report catches new ones as they appear. You answer each question or pass it to the page's owner, and WikiFix updates every page involved. You review every edit before it applies and every edit can be undone in one click.

Examples from NASA, GitLab and PostHog: https://wikifix.ai/examples
Marketplace: https://marketplace.atlassian.com/apps/2836392116

Best,
Rezgar

George Ilie (AIGA Tech)

Hi @Artiom Aleksandrovici ,

On detection, the step before anyone decides, here is what works without an app, and one option with.

Without an app: Confluence search covers every space you can read, so searching a distinctive sentence from the policy in quotes (“within 30 days of purchase”) lists every page carrying it, across spaces. That catches copies, including the one where someone changed 30 to 14 and left the rest. It won’t catch a page that says the same thing in completely different words. CQL title ~ "copy of" also finds copied pages that were never renamed, a common start of two “current” versions. And on ownership: check more than deactivated accounts. Someone removed from Confluence without their account being deactivated still looks like a valid owner.

With an app: Disclosure: I’m the developer of Supersede at AIGA Tech. It fingerprints every page and compares them across all spaces at once, so in your refund example the copies in different spaces come up for review, with AI saying in one sentence whether each pair is a redundant copy or deliberately parallel (one per region, say). It also flags pages whose owner can no longer use Confluence. The limitation is the same as with the search trick: it finds pages that share most of their text, not two independently written pages that disagree.

https://marketplace.atlassian.com/apps/3108262762/supersede

Zoriana Bogutska_Kolekti

Hi @Artiom Aleksandrovici

Rovo could be a great solution for finding pages using certain criteria, e.g. duplicates, outdated information and conflicting instructions. You can try creating an agent that would regularly notify you or page creators about these pages. But you will still need to do manual checks and updates.

My team created a Space Manager feature for Mosaic, that allows you to review all pages across multiple spaces from a single place, filter them by page properties like label, status and owner, and then apply bulk actions (e.g. update, delete, etc.).

If you are interested, I’m happy to run a demo for you to show how it works for your case.

Welcome!

It looks like you're new here. Sign in or register to get started.

Quick Links

Find more posts tagged with

confluence

cloud

confluence-cloud

wiki

confluence-pages