Case study · Media & Publishing
Recovering organic growth with AI content intelligence
A confidential digital publication used AI-assisted content intelligence to diagnose underperforming articles, prioritise editorial fixes and measure improvements against a control group.
Confidential digital publication
- 30%+
- lift in organic clicks for refreshed articles
- 20%+
- improvement in engaged reading time
- 500+
- articles scored for recovery potential
- 70%
- faster editorial diagnosis cycle
The challenge
The publication had a large archive of evergreen articles, but the editorial team could not reliably tell which pieces deserved attention first. Some pages were losing search visibility, some had weak engagement and others were ranking for mismatched intent.
The old process depended on manual spreadsheet reviews and individual editor judgement. It was slow, difficult to repeat and hard to connect to measurable business outcomes. Leadership needed a way to turn the archive into a managed growth portfolio without exposing sensitive traffic, vertical or client-specific details.
The approach
Build an article health model
We combined search, engagement and content signals into a repeatable article health score. Each article was assessed for traffic trend, query mismatch, freshness, internal-link support, engagement depth and content completeness.
Convert diagnosis into editorial briefs
Instead of handing editors another dashboard, the workflow generated plain-language improvement briefs. Each brief explained why an article was underperforming, which fixes were likely to matter and what evidence supported the recommendation.
Find decay
- Flag articles losing impressions, clicks or engagement momentum.
Read intent
- Compare search demand with the article’s actual angle and structure.
Rank fixes
- Prioritise updates by effort, confidence and expected upside.
Measure safely
- Compare refreshed articles with a control group before scaling.
Keep humans in the editorial loop
Editors reviewed every recommendation before publication. The AI layer supplied diagnosis, structure and prioritisation; the editorial team kept final control over voice, standards, sourcing and publication decisions.
The outcome
The publication moved from reactive article updates to a ranked recovery pipeline. Editors could see which pages had the highest upside, why they were underperforming and what to change first.
A measured refresh cohort outperformed the control group across search and engagement signals. The most valuable shift was operational: archive optimisation became a repeatable workflow rather than an occasional audit.
- 30%+
- lift in organic clicks across refreshed articles
- 20%+
- improvement in engaged reading time
- 500+
- articles scored and segmented for action
- 70%
- reduction in manual diagnosis time
The new workflow helped our editors see which articles deserved attention and why, without turning editorial judgement into a black box.
