← back
Case Study · CNET / Ziff Davis · 2025

AI Atlas Redesign

Redesigning CNET’s AI content destination to improve discoverability, deepen engagement, and unlock sponsorship value, under real engineering constraints.

Product Strategy Content Discovery Cross-functional Information Architecture
See AI Atlas live on CNET ↗
01 · Overview

A new destination for AI content

AI Atlas is a CNET initiative that consolidates AI coverage into a single destination and creates a high-value surface for sponsorship. Early performance showed strong traffic, but weak content discovery and shallow engagement, limiting both user value and its sponsorship potential.

Original AI Atlas page before the redesign, with shallow content rivers and low-engagement navigation
Original AI Atlas – before redesign
02 · Problem

Strong traffic. Weak depth.

The data told a clear story: users arrived, skimmed the surface, and left. For a property built around editorial depth and credibility, that was both a UX failure and a business risk to the sponsorship model.

Almost all engagement landed on content rivers, while the navigation meant to guide users (jump links, FAQs, glossary) saw near-zero interaction.

92%
of clicks landed on content rivers – layout misaligned with behaviour
6%
jump-link interaction – core navigation was effectively invisible
53%
of traffic was mobile – the redesign had to work device-agnostically, not desktop-first
03 · My Role

Product Design

I led the redesign across research, editorial, sales, and engineering. Beyond visual execution, I synthesised research into product opportunities, defined the information architecture, and translated competing stakeholder priorities into one coherent direction.

Research synthesis Information architecture UX & prototyping Stakeholder alignment
04 · Insights → Decisions

What the data told us – and what we did about it

Pairing six months of page analytics with 12 unmoderated sessions run on Userlytics – two six-person cohorts, one actively seeking AI content and one merely likely to encounter it – surfaced consistent patterns. Each one mapped directly to a product decision.

Insight
Users engage with content rivers, not structured navigation
Decision
Prioritise content rivers at the top instead of fighting behaviour
Insight
High-intent users want a fast path to a specific content type
Decision
Introduce a filterable directory built for retrieval
Insight
Jump links and FAQ modules were ignored across sessions
Decision
Make navigation persistent and contextually visible
Insight
Editorial credibility drove trust in AI content
Decision
Elevate authors and expertise as first-class UI
Insight
A majority of traffic was mobile, but the page was built desktop-first
Decision
Rebuild the core navigation to be persistent and device-agnostic: sticky jump-links and a filterable directory that hold up on small screens
05 · Visual Direction

Less robotic, more human

Testing surfaced two brand problems: users felt overstimulated by the page’s colours and visuals, and the palette read as overly feminine for the broader audience we needed to reach. This was my first project owning branding end-to-end, and I anchored the refresh in three moves.

01 · Palette

Pull back the colour

Removed the light purple and reduced overall colour so the brand reads more sophisticated and digital – and less overstimulating.

02 · Imagery

Add human imagery

More people, fewer abstractions – so users can envision themselves using AI and the page feels relatable rather than robotic.

03 · Motion

Lean into animation

Motion graphics (built with our motion design team) make the hub feel modern and guide users down the page.

06 · Product Strategy

One destination, two surfaces

The core insight: AI Atlas was serving two user modes (passive browsing and active search) from one experience, and doing neither well. I split it into a dual-surface architecture that separated the modes without fragmenting the destination.

Redesigned AI Atlas hub as shipped at cnet.com/ai-atlas, with content rivers prioritized above persistent navigation and author credibility elevated
AI Atlas Hub – the redesigned discovery surface, live at cnet.com/ai-atlas
Redesigned AI Atlas directory, filterable by topic, type, and author for high-intent search
AI Atlas Directory – filterable intent surface

AI Atlas Hub

Discovery · passive browsing
Content rivers ordered by engagement data
Persistent navigation across sections
Video and social woven in to grow dwell time
Author credibility elevated throughout

AI Atlas Directory

Intent · active search
Filterable by topic, type, and author
Simplified layout tuned for retrieval
Clear pathways back into the hub
Structured to scale as coverage grows
07 · Tradeoffs & Execution

Phasing the launch under constraints

Engineering capacity meant we couldn’t ship both surfaces at once. Rather than delay the whole redesign, I phased it – shipping the Hub first to start generating real signal, with the Directory following.

V1 · Shipped

Hub experience

Full hub redesign with river prioritisation, persistent navigation, and elevated editorial credibility.

V2 · Shipped later

Directory

Filterable directory with cross-surface navigation. Both surfaces now live at cnet.com/ai-atlas.

08 · Impact

What shipped – and what it moved

📈
Jump-link usage 6% → 20%
Navigation engagement more than tripled after the Hub launch.
⏱️
Time on site up ~10%
Dwell time grew within six months of launch.
🤝
2 sponsorships pre-launch
Closed on prototypes before a line of code shipped.
🚀
Both surfaces live
The full dual-surface experience is live.
cnet.com/ai-atlas →
09 · Reflection

What I’d do differently

The biggest gap was entering without explicit, measurable success metrics tied to each decision. We had clear qualitative goals, but not the quantitative targets that make post-launch evaluation rigorous – a PM habit I’ve since adopted.

I’d also instrument prototypes earlier. Two sponsorships closing on prototypes alone signalled stronger commercial demand than we realised; testing with sales sooner could have shaped strategy from the start.