K.

Back

K.

Back

K.

AI Powered Personalisation

Building a personalised homepage for the Radio Times app, grounded in research on how users connect content.

Project details

Team

Kanza Leghari

Snr Product Designer

Magda G.

Product Designer

Software Engineers

Alex M.

Lead App Developer

Platforms & Softwares

UserTesting

User research & Card sorting

Figma

Design & Prototyping

Figma + Claude

Claude & FAISS

Vector embeddings

Summary

Users across IM native apps, particularly Radio Times, often reached dead-ends in their browsing journeys, with related content relying on basic rules or keyword matching that made recommendations feel generic. Journey mapping, card sorting and user research showed that users connect content through shared tone, genre and cast, and these insights shaped a new recommendation model built on vector embeddings.

I led the design of For You, an alternative homepage in the Radio Times app that recommends films and shows based on each user’s behaviour and stated preferences. In an A/B test against the editorial homepage, For You delivered a statistically significant uplift in click-through rate and has since been rolled out to all users. The same model also powers Film Finder, a mood-based discovery flow elsewhere in the app.

Users across IM native apps, particularly Radio Times, often reached dead-ends in their browsing journeys, with related content relying on basic rules or keyword matching that made recommendations feel generic. Journey mapping, card sorting and user research showed that users connect content through shared tone, genre and cast, and these insights shaped a new recommendation model built on vector embeddings.

I led the design of For You, an alternative homepage in the Radio Times app that recommends films and shows based on each user’s behaviour and stated preferences. In an A/B test against the editorial homepage, For You delivered a statistically significant uplift in click-through rate and has since been rolled out to all users. The same model also powers Film Finder, a mood-based discovery flow elsewhere in the app.

Increase in stickiness (A/B test results)

Uplift in session durations

Continue reading for the detailed version.

Research &
Discovery

Flow mappings

Mapping key journeys across our native apps, particularly Radio Times, showed where discovery was breaking down. Users could reach content easily, but the onward journey was often unsupported, with low-quality related content and no clear next step.

Card sorting and user research

To understand what meaningful next steps look like, we explored how users expect content to relate. They consistently preferred recommendations that felt contextually relevant, connected through relationships such as franchise, genre, sub-genre, tone and cast, rather than simple metadata or keyword overlap.

From rules-based to AI-powered

To enable more meaningful recommendations, the team explored an alternative to building increasingly complex search queries. Based on the research, users were looking for content that felt similar in tone, mood or narrative, something that traditional rules-based systems struggled to capture.

Instead, we moved towards an AI-driven approach that represents each piece of content using multiple signals, including titles, genres and descriptions. By placing greater emphasis on descriptive context, the system can identify deeper relationships between content, such as shared tone or themes, rather than relying on keyword overlap alone.

This allows recommendations to be generated based on overall similarity, surfacing content that feels genuinely related in experience. For example, rather than matching on a single keyword, the system can identify films or shows that share a similar mood, storyline or genre blend, resulting in more relevant and intuitive recommendations.

Design & Experimentation

Phase 2

Designing For You

With the model in place, I designed For You as an alternative homepage that puts personalised recommendations front and centre. The page draws on two kinds of signal. Behavioural signals include what a user has added to their watchlist, what they’ve marked as watched and the genres they tend to watch trailers for. Stated preferences include the streaming providers they subscribe to and any linear TV channels they’ve favourited.

These signals populate carousels such as Top Picks for You. Users can switch between All, Films and TV Shows, filter the page by a specific streaming provider, and update their providers directly from the page if their subscriptions change.

User Testing: Rapid prototyping with Claude Design

To move quickly from concept to testing, I used Claude Design to build interactive prototypes of For You. This let me explore layout and filtering options in a realistic, clickable format and put them in front of users early, so feedback could shape the design before it reached development.

Knowing that many Radio Times users still rely on linear TV, I added an On TV Tonight carousel to the page. User testing validated this, with participants responding well to seeing personalised picks from live TV alongside streaming recommendations, so it became part of the final design.

A/B testing against the editorial homepage

To understand whether a personalised page would outperform the curated one, we ran an A/B test where users landed on either the existing editorial homepage or For You. For You showed a statistically significant uplift in click-through rate compared to the editorial homepage, and we have since rolled it out to all users.

Outcomes &
Learnings

For You is now set to be live soon for all Radio Times app users after delivering a statistically significant uplift in click-through rate over certain the editorial homepage carousels. The same model also powers Film Finder, a mood-based discovery flow covered in its own case study, and we are now exploring similar personalised experiences across our other native apps.

This project reinforced the value of using AI to meet genuine user needs rather than adding it as a feature in itself. Defining the right signals upfront, and validating ideas like On TV Tonight early through rapid prototyping, kept the design grounded in how our audience actually watches.

For You is now set to be live soon for all Radio Times app users after delivering a statistically significant uplift in click-through rate over certain the editorial homepage carousels. The same model also powers Film Finder, a mood-based discovery flow covered in its own case study, and we are now exploring similar personalised experiences across our other native apps.

This project reinforced the value of using AI to meet genuine user needs rather than adding it as a feature in itself. Defining the right signals upfront, and validating ideas like On TV Tonight early through rapid prototyping, kept the design grounded in how our audience actually watches.