AI-Ready Score
The AI-Ready Score is a 0–100 measure of how usable a page is for AI tools — Atlassian Rovo and any large-language-model integrations that index your Confluence content. A page can be Healthy but poorly AI-Ready if it lacks structure that AI systems lean on to understand it.
Why a separate score?
Section titled “Why a separate score?”Health and AI-readiness overlap but are not the same thing.
- A Healthy page can be a single long paragraph that humans read fine but AI tools struggle to chunk.
- An AI-Ready page is one that’s structured with clear headings, working links, a real title, and enough depth for an AI to draw context.
Tracking them separately lets you see which pages humans will love and which AI assistants will be able to use confidently.
Where the score shows up
Section titled “Where the score shows up”- Global Dashboard — average AI-Ready across all pages.
- Space Dashboard — average AI-Ready for that space and a “Not AI-Ready” count card.
- All Pages tab — every page’s individual AI-Ready score, sortable.
- Inline byline — the AI-Ready value next to the Health Score.
What the score is built from
Section titled “What the score is built from”Like the Health Score, AI-Ready combines several signals (visible by name in the Settings panel). The weighting differs — it’s tuned for how AI systems consume content rather than how humans skim it.
Signal categories used for AI-Ready:
- Structure — well-formed heading hierarchy is the biggest single contributor. Pages without headings score low even if everything else is good.
- Content depth — there has to be something substantive to draw from; short stubs aren’t useful context.
- Freshness — older pages are more likely to be inaccurate, so recency contributes here too.
- Clarity — placeholder titles, “untitled” pages, and pages with almost no content reduce the score.
- Connectivity — pages with working internal links score higher; broken links are not counted as connectivity.
The exact weights and thresholds are part of Bytera’s content model and tuned over time.
Score bands
Section titled “Score bands”| Band | What it usually means |
|---|---|
| AI-Ready | The page has the structure and substance an AI assistant needs to answer questions from it. |
| Improvable | One or two structural signals are missing — likely headings or connectivity. |
| Not AI-Ready | The page is too thin, untitled, or unstructured for AI tools to use confidently. |
The Space Dashboard surfaces a Not AI-Ready count card so you can spot how much of a space needs structural work.
Improving a page’s AI-Ready Score
Section titled “Improving a page’s AI-Ready Score”Roughly in order of impact:
- Add a clear heading structure — a top-level page heading then 2–3 supporting H2/H3 sections.
- Use a real title — not “Untitled page”, not a placeholder.
- Write enough. A one-line page can’t be made AI-ready; expand it to a few paragraphs.
- Fix broken links — broken links don’t count as connectivity.
- Add a couple of internal links to related pages, so the AI can see what this page connects to.
- Keep it current — a recently-edited page scores higher in freshness.
These are the same moves that lift the Health Score; AI-Ready just weights Structure more heavily.
Why “AI-Ready” matters for Atlassian Rovo
Section titled “Why “AI-Ready” matters for Atlassian Rovo”Atlassian Rovo and similar AI tools index Confluence pages and surface them in chat-style answers. The quality of those answers depends heavily on the structure of the underlying content. A team that brings its AI-Ready average up generally sees Rovo answers improve in the same week — there’s no model change, just better source material.
- Content Health Score — the broader companion score.
- Custom Policies — adjust the rules that feed into scoring.
- Trends — watch AI-Ready improve over time.