> ## Documentation Index
> Fetch the complete documentation index at: https://user-guides.demand-genius.com/llms.txt
> Use this file to discover all available pages before exploring further.
# How Content Intelligence works
> How your site is crawled, tagged and scored: the Library grid, the five KPIs and twelve signals, page status, refresh cadence, Insights and Tree.
Content Intelligence is a qualitative content audit run at scale. Every page the crawler finds on your site is read by AI, tagged by what it is and who it is for, and scored against the things that decide whether an AI assistant is likely to pick it up, quote it and recommend you off the back of it. Those scores refresh on a schedule, so a subjective judgement about content quality becomes a KPI you can track and improve.
AI Search Intelligence tells you how AI talks about you today. Content Intelligence tells you how you are influencing that, where the gaps are and which pages to work on first.
## The Library
Open **Library** under Content Intelligence. This is a database of all your content, which our AI indexes and evaluates. Each row is a page on your site; each column is something the platform knows about it from our AI analysis.
### Standard columns
| Column | What it holds |
| ------------------------------------------------------------------------------------------------ | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| **Title**, **URL** | The page |
| **Scrapable** | Deselect this to stop us analysing that page. |
| **Page Status** | Indexed, noindex, 301, 404 or 503, from a status probe on each crawl |
| **Content Type**, **Topic**, **Funnel Stage**, **Industry**, **Page Type**, **Primary Audience** | AI classification tags, editable |
| **Marketing Content** | Whether the page is marketing content or utility (legal, login, careers). Useful for filtering. |
| **Language**, **Word Count**, **Date Published**, **Date Modified** | Read from the page - this pulls directly from the metadata. If no published or modified date is shown, this is something to update in your CMS. |
| **Description**, **Summary** | Meta description and an AI summary |
| **Confidence**, **Explanation** | How sure the classifier was and why |
| **Quality, Extractability, Citation Likelihood, Strategic Differentiation, Accuracy & Currency** | Subjective, AI analysis of critical content attributes that drive AI-era performance. Each is composed of 2-4 specific attributes, which we roll up to a score. Expand the column to see the underlying analysis. |
Tag values can be edited inline and in bulk for selected rows. **Manage Columns** shows and hides columns and sets their order. **Export CSV** downloads everything or the filtered view.
### KPI column groups
Five collapsible column groups hold the content scores. Collapsed, each shows a score from 0 to 100, its band, the change since the previous capture and a sparkline. Expanded, it shows the underlying signals as labels.
| KPI | Signals |
| ----------------------------- | --------------------------------------------------------- |
| **Content Quality** | Definition Clarity, Content Depth, Keyword Optimisation |
| **AI Extractability** | Content Structure, AI Extractability |
| **Citation Likelihood** | Brand Assertion, Authority Signals |
| **Strategic Differentiation** | Narrative Ownership, Information Gain |
| **Accuracy & Currency** | Information Currency, Product Mention, Product Validation |
## How scores are produced
Each page is read in full and each of the twelve signals is given a three-level label by the model, using a fixed rubric. Labels map to 2, 1 or 0. A KPI's score is the sum of its signal values divided by the maximum possible, as a percentage. A signal the model could not assess is left out of both sides rather than guessed.
| Band | Score |
| -------------- | ------------ |
| **Strong** | 67 and above |
| **Developing** | 45 to 66 |
| **At risk** | below 45 |
An **Overall** bar above the grid summarises the library. Because the scoring is done by AI it is directional rather than exact: good enough to justify a recommendation and to show that work moved something, not a laboratory measurement.
If the KPI cells show an **Off** chip, KPI scoring has not been switched on for your workspace. Ask Demand-Genius to enable it ([tom@demand-genius.com](mailto:tom@demand-genius.com)). A dash means the page has not been scored yet; a zero is a real zero.
## When things refresh
Two schedules run independently.
**Content discovery.** The crawler re-reads your sitemap on the auto-update day set for the library (see [Directories](/content-intelligence/directories)), adds new pages and re-checks page status.
**KPI scoring.** A nightly job scores pages that are due. How often a page is re-scored depends on your plan: some plans score each page once, others quarterly or monthly. A per-page **Re-run** action scores a single page on demand, limited to once every five minutes per page and thirty pages an hour per workspace. This is intended to let you kick off single-page rescores where you feel the AI may have hallucinated, or have updated a page and want to see if it's improved the score, and is subject to fair usage. Scoring history is kept, which is what the deltas and trendlines draw on.
## Insights
**Insights** charts the library across its tags: **Content Volume by Type**, **Content Volume by Topic**, **Publishing Frequency** (how often you publish new content) and **Modified Frequency** (how often you refresh existing content), plus a heatmap you can pivot by Topic, Content Type, Industry, Audience, Funnel Stage or any Smart Column. This helps you see where your authority lies, and can help to contextualise AI visibility trends and identify opportunities.
## Tree
**Tree** shows the same pages as a site hierarchy, so you can see where content sits and expand or collapse branches. It is a navigation aid rather than a report.
## Performance columns
If Performance is enabled for your workspace, the grid also carries a **Performance** group: **Traffic Total** and **Influenced MQL / SQL / Closed Won** per page, fed by the tracking script and your CRM. See [Performance tracking](/content-intelligence/performance-tracking).
## Next
Run your own analysis across the library.
Traffic and pipeline influence per page.
Add competitor libraries and set update schedules.
Missing pages, blocked crawlers, Shopify.