ShapersOfAI.com · From Users to Shapers · AI Leadership · Future of Work · AI Skills

AI Use Case

Intranet Content Quality Dashboard

Scores intranet content and shows editors which pages need improvement, helping keep information clear, current and usable by employees, enterprise search and AI assistants.

Intranets often contain thousands of pages created by different editors over many years. Content becomes outdated, incomplete or difficult to understand, while responsibility for improving it remains unclear.

An Intranet Content Quality Dashboard continuously or regularly evaluates pages against defined editorial criteria. It can assess elements such as clarity, structure, completeness, metadata, freshness and ownership. Results are presented through an accessible red, yellow and green rating, allowing editors and content owners to see immediately where action is required.

The assessment can also consider whether content provides enough context for enterprise search and AI assistants such as Microsoft Copilot to interpret and reuse it correctly. In a more advanced version, AI generates concrete recommendations for improving each page. Editors remain responsible for reviewing and implementing these suggestions.

Painpoints

Large volumes of intranet content are difficult to review manually
Outdated or incomplete pages remain online
Content ownership is unclear
Editorial quality varies between departments and editors
Editors lack a shared definition of good intranet content
Poorly structured pages are difficult for employees to find and understand
Enterprise search and AI assistants may return incomplete or misleading answers
Content problems often become visible only after employees complain
Central communication teams cannot review every page themselves

Goals

Make content quality visible across the intranet
Give editors clear responsibility for their pages
Identify outdated, incomplete or poorly structured content
Establish consistent editorial standards
Prioritise pages that require attention
Improve findability and comprehension for employees
Make content easier for enterprise search and AI assistants to interpret
Reduce the manual workload of central intranet teams
Turn quality assessment into concrete editorial action

Risks

Automated scores may oversimplify editorial quality
Editors may optimise for the score instead of the reader
Incorrect rules could classify good content as problematic
AI-generated recommendations may change facts or intended meaning
Pages with restricted or sensitive information require careful processing
Editors may distrust ratings they cannot understand
Outdated assessment criteria can produce misleading results
The dashboard may identify problems without ensuring that anyone resolves them

Implementation Idea

Define a concise editorial quality model with approximately 8–12 measurable criteria
Include criteria such as title clarity, page summary, heading structure, readability, completeness, metadata, content owner and review date
Add AI-readiness criteria such as explicit context, unambiguous terminology and self-contained information
Assign each criterion a transparent weighting
Connect the dashboard to the intranet or SharePoint content inventory
Read page content, metadata, ownership and last-review information
Use rules for objectively measurable criteria such as missing owners, dates, summaries or headings
Use an LLM only for qualitative assessments such as clarity, comprehensibility and missing context
Calculate an overall red, yellow or green status for every page
Show editors exactly which criteria caused the rating
Provide filtered views by department, content owner, page type and status
Send editors a regular overview of pages requiring attention
Generate concrete improvement recommendations for red and yellow pages
Let editors accept, modify or reject each recommendation
Never publish AI-generated changes automatically
Record when a page was reviewed and when the score was last calculated
Reassess updated pages to show whether their score improved
Regularly compare automated ratings with human editorial reviews and refine the criteria

Practical Microsoft implementation

A first version could combine:

SharePoint as the content source
SharePoint Lists or Dataverse for scores, criteria and review status
Power Automate to trigger recurring assessments and notify content owners
Azure OpenAI or an approved Microsoft AI service for qualitative content analysis
Power BI for the red-yellow-green dashboard
Copilot Studio for an optional assistant that explains ratings and recommends improvements

A simpler pilot does not need the complete dashboard. Start with 50–100 representative intranet pages, export their content and metadata, apply the quality model, and compare the results with assessments by experienced editors.

Advatera Community Examples

Allianz is developing a dashboard that helps intranet editors monitor the quality of their content. Pages receive an accessible red, yellow or green assessment, making content quality visible and encouraging editors to take responsibility for their pages.

A planned next step is to add AI-powered recommendations that show editors how individual pages can be improved. The approach is especially relevant as intranet content must work not only for employees but also as a reliable source for enterprise search and AI assistants.