BACK TO HOME

NEXT PROJECT /02

Neutrality.inc

Neutrality

A research-intelligence platform for journalists, analysts, and researchers. Discovery, synthesis, and publication in one place.

Live Preview

Role

UX UI Designer

Duration

20 months

Research Intelligence / Saas

Type

End-to-end platform redesign

Context

From information discovery to research intelligence

Neutrality was built to help people make sense of complex topics by connecting discovery, intelligence layers, collections, research workspaces, topic-building, and publishing.

Discovery, synthesis, and publication without losing context, trust, or signal.

But when I joined, the vision had evolved faster than the experience. Users could search, save, build, and share, but these actions felt like separate tools rather than one connected research journey.

"The opportunity was to turn Neutrality from a collection of features into a coherent research intelligence platform"

The Problem

The product was functional, but not coherent

Neutrality did not fail because it lacked features. It failed because the features did not compound into a clear research workflow.

Each surface solved a small part of the journey, but the system did not help users move from one part to the next.

Neutrality did not fail because it lacked features. It failed because the features did not compound into a clear research workflow.

Each surface solved a small part of the journey, but the system did not help users move from one part to the next.

Search helped users find articles, but most sessions ended there. Collections allowed users to save content, but users treated them like folders and rarely returned.

The deeper problem was continuity.

" Users were not building research over time. They were repeatedly starting again."

Topic Builder existed as a creation tool, but users did not understand whether it was for writing, arranging, publishing, or analysing. Dashboard showed activity, but did not help returning users regain context.

The deeper problem was continuity.

" Users were not building research over time. They were repeatedly starting again."

My Approach

Continuous product learning, not a one-time redesign..

Discovery helped define the larger product problem: Neutrality was not just fragmented visually, it lacked continuity across the research journey.

Each cycle followed the same rhythm

Discover → Define → Develop → Deliver

Each launch also started with a clear Job To Be Done, so the work stayed focused on user behaviour instead of just feature delivery.

D1 clarified the product direction.
CL1 applied the loop to Search + Collections.
CL2 applied it to Topic Builder + Dashboard.
CL3 applied it to Hubs, Profiles, and Publisher Layer.

The design system evolved underneath all launches, making each cycle stronger instead of adding new fragmentation.

Discovery helped define the larger product problem: Neutrality was not just fragmented visually, it lacked continuity across the research journey.

Each cycle followed the same rhythm

Discover → Define → Develop → Deliver

Each launch also started with a clear Job To Be Done,

So the work stayed focused on user behaviour instead of just feature delivery.

D1 clarified the product direction.
CL1 applied the loop to Search + Collections.
CL2 applied it to Topic Builder + Dashboard.
CL3 applied it to Hubs, Profiles, and Publisher Layer.

The design system evolved underneath all launches, making each cycle stronger instead of adding new fragmentation.

Research Synthesis

4C framework - Finding what was broken

I used the 4C framework — Collect, Choose, Create, Commit to make sense of four evidence streams:

I used the 4C framework — Collect, Choose, Create, Commit to make sense of four evidence streams:

Across frown responses, interviews, heuristic review, and competitor analysis, the same issues kept repeating.

Across frown responses, interviews, heuristic review, and competitor analysis, the same issues kept repeating.

Research continuity was broken
Users collected information but struggled to synthesize it
The product conflicted with users’ mental models
Trust and credibility were not clear enough

Research continuity was broken

Users collected information but struggled to synthesize it
The product conflicted with users’ mental models
Trust and credibility were not clear enough

Before moving to next phase, three strategic decisions were locked:

Primary user - Journalists, analysts, and research students
Product identity - Research intelligence platform
Problem type - System-level continuity failure

This reframed the project from fixing isolated screens to solving a system-level problem.

Neutrality needed to help research accumulate across sessions, surfaces, and outputs.

This led to four design principles

Two deeper issues were driving most of the friction:

Cognitive overload — Users were dealing with too much information, too early, without enough structure.
Credibility breakdown — Users needed to understand why information should be trusted, not just see a score or badge.

A user could click more, save more, and spend more time in the platform while still feeling confused. For Neutrality, success meant something more specific. Engagement alone was not the right success metric.

North Star Metric

" User confidence in neutral understanding of complex topics "

North Star Metric —

" User confidence in neutral understanding of complex topics "

North Star Metric

" User confidence in neutral understanding of complex topics "
The research moved the project from feature fixes to system redesign.

The research moved the project from feature fixes to system redesign.

Design Strategy

Three launches, one connected system

The redesign was organised into three core launches.

This sequencing was important because each launch solved one layer of the product problem and revealed what the next layer needed to address.

So the strategy followed the research journey. The sequencing created a platform logic:

`Discovery → Workspace → Synthesis → Orientation → Publication → Trust`

Each launch solved the next layer of the research journey.

The redesign was organised into three core launches.

This sequencing was important because each launch solved one layer of the product problem and revealed what the next layer needed to address.

So the strategy followed the research journey. The sequencing created a platform logic:

`Discovery → Workspace → Synthesis → Orientation → Publication → Trust`

Each launch solved the next layer of the research journey.

Core Launch 1 — Search + Collections

The goal of CL1 was to change that behaviour.

Search needed to become a transition into research.

Search solution
The redesigned Search followed one principle: Clean results first. Intelligence on demand.
Collections needed to become workspaces.

Collections solution
Collections were redesigned into four purpose-led tabs. The principle was: Purpose visible through placement, not labelling.

The goal of CL1 was to change that behaviour.

Search needed to become a transition into research.
Search solution
The redesigned Search followed one principle: Clean results first. Intelligence on demand.
Collections needed to become workspaces.
Collections solution
Collections were redesigned into four purpose-led tabs. The principle was: Purpose visible through placement, not labelling.

Users no longer needed to explore every option to understand the workspace.

Turning discovery into active research

Core Launch 2 — Topic Builder + Dashboard

CL1 helped users find and organise research. But the next problem was synthesis. CL2 solved both sides of the problem. Researchers needed to arrange evidence before forming conclusions.

The key insight was:
1. A document editor asks, “What do you want to say?”
2. A research canvas asks, “What patterns do you see?”

CL1 helped users find and organise research. But the next problem was synthesis. A Collection link gave someone a list of files, not a narrative. Dashboard did not show what mattered, what had changed, or where to resume.

CL2 solved both sides of the problem.

Topic Builder problem

Topic Builder was originally planned as a rich text document editor. Researchers needed to arrange evidence before forming conclusions.

Topic Builder solution

The redesigned Topic Builder became a flexible panel canvas.

Three decisions made the model work:

Assets integration — Saved Topics flowed from Search into Topic Builder
Publishing clarity — Preview, Publish, and Last Synced states stayed visible

Shared panel language — Every panel used the same controls and behaviour

The key insight was:
1. A document editor asks, “What do you want to say?”
2. A research canvas asks, “What patterns do you see?”

Topic Builder solution

The redesigned Topic Builder became a flexible panel canvas.

Three decisions made the model work:

Assets integration — Saved Topics flowed from Search into Topic Builder
Publishing clarity — Preview, Publish, and Last Synced states stayed visible

Shared panel language — Every panel used the same controls and behaviour

Topic Builder

The canvas gave users freedom, while the shared panel system kept the experience coherent.

Dashboard redesign focused on one question:

What does a returning researcher need to know in the first 10 seconds?

Dashboard existed, but it did not help returning users. They had to remember what they were working on, what had changed, and where to continue.

The redesign focused on one question:

What does a returning researcher need to know in the first 10 seconds?

Dashboard solution

Dashboard became a research command centre.

Dashboard solution

Dashboard became a research command centre.

It summarised the state of active work across three panel types:

1. Saved Topics — Article count, last activity, research momentum
2. Topic Builder files — Publish status, last synced, collaborators

Collections — Items, shared users, last modified

It summarised the state of active work across three panel types:

1. Saved Topics — Article count, last activity, research momentum
2. Topic Builder files — Publish status, last synced, collaborators

Collections — Items, shared users, last modified

The design rationale was:

" The Dashboard remembers research intent on behalf of users."

The design rationale was:

" The Dashboard remembers research intent on behalf of users."

Dashboard

The Available Assets sidebar kept Saved Topics and Collections within reach.

Core Launch 3
Publisher Layer.
Hub + Profile.

Designing trust as infrastructure

After CL2, users could create and publish research. But publishing introduced a new problem: trust. Readers could access published work, but they could not easily judge who created it, what supported it, or how credible it was.

The problem was that published research needed credibility context.

CL3 introduced three connected layers: Publisher Layer, Hubs, Profiles

Publisher Layer

The Publisher Layer supported external content management and campaigns.

Publishers could manage RSS feeds, external content, and promoted research. But paid distribution did not remove credibility requirements.

The same trust signals applied to paid and free content.

Publisher Layer

The Publisher Layer supported external content management and campaigns.

Publishers could manage RSS feeds, external content, and promoted research. But paid distribution did not remove credibility requirements.

The same trust signals applied to paid and free content.

Publisher Layer

Hubs

Hubs became trust-led discovery surfaces.

Published Collections and Topic Builder files appeared with credibility context, not just thumbnails or titles.

Profiles

Profiles were elevated from account settings into public trust anchors.

Two profile types were designed:

Individual profile — Shows researcher identity, contribution history, published work, and credibility indicators
Institution profile — Shows organisational credibility and published research context

Hubs

Hubs became trust-led discovery surfaces.

Published Collections and Topic Builder files appeared with credibility context, not just thumbnails or titles.

Profiles

Profiles were elevated from account settings into public trust anchors.

Two profile types were designed:

Individual profile — Shows researcher identity, contribution history, published work, and credibility indicators
Institution profile — Shows organisational credibility and published research context

" Trust became part of the system, not a badge on top of it."

Hubs

Profiles

Final Output

" Research is no longer just something you do. It's something you own, present, and profit from."

Search

Not a search engine. A research engine.

Search

Not a search engine. A research engine.

Platform impact

The final platform helped users move from finding information to building research, publishing outputs, and evaluating credibility.

From 64% single-session drop-off to 85% multi-session continuation. Neutrality didn't just improve research, it changed how it happens.

From 64% single-session drop-off to 85% multi-session continuation. Neutrality didn't just improve research — it changed how it happens.

Platform scale

Research articles analysed

Structured collections created

Multi-session continuation rate

Behaviour change

Research depth

Collections adoption

Topic Builder engagement

Trust-led navigation

Observed after 1 month of beta launch. Validated through moderated behavioural analytics.

Platform scale

Research articles analysed

Structured collections created

Multi-session continuation rate

Behaviour change

Research depth

Collections adoption

Topic Builder engagement

Trust-led navigation

Observed after 1 month of beta launch. Validated through moderated behavioural analytics.

Platform scale

Research articles analysed

Structured collections created

Multi-session continuation rate

Behaviour change

Research depth

Collections adoption

Topic Builder engagement

Trust-led navigation

Observed after 1 month of beta launch. Validated through moderated behavioural analytics.

★ CURRENTLY OPEN TO WORK ★ BASED IN SCOTLAND, UK ★ UX · UI · RESEARCH · VIBECODE
★ CURRENTLY OPEN TO WORK ★ BASED IN SCOTLAND, UK ★ UX · UI · RESEARCH · VIBECODE

Neutrality.inc

Neutrality

Neutrality

A research-intelligence platform for journalists, analysts, and researchers. Discovery, synthesis, and publication in one place.

A research-intelligence platform for journalists, analysts, and researchers. Discovery, synthesis, and publication in one place.

Role

UX UI Designer

Duration

20 months

Category

Research Intelligence / Saas

Type

End-to-end platform redesign

Context

Context

From information discovery to research intelligence

From information discovery to research intelligence

Neutrality was built to help people make sense of complex topics by connecting discovery, intelligence layers, collections, research workspaces, topic-building, and publishing.

Discovery, synthesis, and publication without losing context, trust, or signal.

But when I joined, the vision had evolved faster than the experience. Users could search, save, build, and share, but these actions felt like separate tools rather than one connected research journey.

"The opportunity was to turn Neutrality from a collection of features into a coherent research intelligence platform"

The Problem

The Problem

The product was functional, but not coherent

The product was functional, but not coherent

Neutrality did not fail because it lacked features. It failed because the features did not compound into a clear research workflow.

Each surface solved a small part of the journey, but the system did not help users move from one part to the next.

Neutrality did not fail because it lacked features. It failed because the features did not compound into a clear research workflow.

Each surface solved a small part of the journey, but the system did not help users move from one part to the next. Search helped users find articles, but most sessions ended there. Collections allowed users to save content, but users treated them like folders and rarely returned.

The deeper problem was continuity.

" Users were not building research over time. They were repeatedly starting again."

Topic Builder existed as a creation tool, but users did not understand whether it was for writing, arranging, publishing, or analysing. Dashboard showed activity, but did not help returning users regain context.

The deeper problem was continuity.

" Users were not building research over time. They were repeatedly starting again."

My Approach

My Approach

Continuous product learning, not a one-time redesign..

Continuous product learning, not a one-time redesign..

Discovery helped define the larger product problem: Neutrality was not just fragmented visually, it lacked continuity across the research journey.

Each cycle followed the same rhythm

Discover → Define → Develop → Deliver

Each launch also started with a clear Job To Be Done, so the work stayed focused on user behaviour instead of just feature delivery.

D1 clarified the product direction.CL1 applied the loop to Search + Collections.CL2 applied it to Topic Builder + Dashboard.CL3 applied it to Hubs, Profiles, and Publisher Layer.

The design system evolved underneath all launches, making each cycle stronger instead of adding new fragmentation.

Discovery helped define the larger product problem: Neutrality was not just fragmented visually, it lacked continuity across the research journey.

Each cycle followed the same rhythm

Discover → Define → Develop → DeliverEach launch also started with a clear Job To Be Done,

So the work stayed focused on user behaviour instead of just feature delivery.

D1 clarified the product direction.CL1 applied the loop to Search + Collections.CL2 applied it to Topic Builder + Dashboard.CL3 applied it to Hubs, Profiles, and Publisher Layer.

The design system evolved underneath all launches, making each cycle stronger instead of adding new fragmentation.

Research Synthesis

Research Synthesis

4C framework - Finding what was broken

4C framework - Finding what was broken

I used the 4C framework — Collect, Choose, Create, Commit to make sense of four evidence streams:

I used the 4C framework — Collect, Choose, Create, Commit to make sense of four evidence streams:Across frown responses, interviews, heuristic review, and competitor analysis, the same issues kept repeating.

Across frown responses, interviews, heuristic review, and competitor analysis, the same issues kept repeating.

Research continuity was broken

Users collected information but struggled to synthesize it

The product conflicted with users’ mental models

Trust and credibility were not clear enough

Research continuity was broken

Users collected information but struggled to synthesize it

The product conflicted with users’ mental models

Trust and credibility were not clear enough

Before moving to next phase, three strategic decisions were locked:

Primary user - Journalists, analysts, and research students

Product identity - Research intelligence platform

Problem type - System-level continuity failure

This reframed the project from fixing isolated screens to solving a system-level problem.Neutrality needed to help research accumulate across sessions, surfaces, and outputs.

This led to four design principles

Two deeper issues were driving most of the friction:

Cognitive overload — Users were dealing with too much information, too early, without enough structure.

Credibility breakdown — Users needed to understand why information should be trusted, not just see a score or badge.

A user could click more, save more, and spend more time in the platform while still feeling confused. For Neutrality, success meant something more specific. Engagement alone was not the right success metric.

North Star Metric

" User confidence in neutral understanding of complex topics "

North Star Metric —

" User confidence in neutral understanding of complex topics "

North Star Metric

Each surface solved a small part of the journey, but the system did not help users move from one part to the next.

Search helped users find articles, but most sessions ended there. Collections allowed users to save content, but users treated them like folders and rarely returned.

D1 clarified the product direction.
CL1 applied the loop to Search + Collections.
CL2 applied it to Topic Builder + Dashboard.
CL3 applied it to Hubs, Profiles, and Publisher Layer.

Discover → Define → Develop → Deliver

Each launch also started with a clear Job To Be Done,

D1 clarified the product direction.
CL1 applied the loop to Search + Collections.
CL2 applied it to Topic Builder + Dashboard.
CL3 applied it to Hubs, Profiles, and Publisher Layer.

I used the 4C framework — Collect, Choose, Create, Commit to make sense of four evidence streams:

Across frown responses, interviews, heuristic review, and competitor analysis, the same issues kept repeating.

This reframed the project from fixing isolated screens to solving a system-level problem.

Neutrality needed to help research accumulate across sessions, surfaces, and outputs.

" User confidence in neutral understanding of complex topics "
The research moved the project from feature fixes to system redesign.

`Discovery → Workspace → Synthesis → Orientation → Publication → Trust`

`Discovery → Workspace → Synthesis → Orientation → Publication → Trust`

Search needed to become a transition into research.

Search solution

Search needed to become a transition into research.
Search solution

The key insight was:
1. A document editor asks, “What do you want to say?”
2. A research canvas asks, “What patterns do you see?”

CL1 helped users find and organise research. But the next problem was synthesis. A Collection link gave someone a list of files, not a narrative. Dashboard did not show what mattered, what had changed, or where to resume.

CL2 solved both sides of the problem.

The redesigned Topic Builder became a flexible panel canvas.

Three decisions made the model work:

The key insight was:
1. A document editor asks, “What do you want to say?”
2. A research canvas asks, “What patterns do you see?”

Topic Builder solution

The redesigned Topic Builder became a flexible panel canvas.

Three decisions made the model work: