Purl AI

Purl’s Kanban board on a laptop screen

Project overview

Co-founded and designed an AI-powered ideation platform. Chat interfaces make you work one idea at a time and lose the rest to scroll. Purl replaced that with a board, so marketing teams could hold several ideas open at once, develop them in parallel, and carry the good ones through to a critiqued, actionable strategy.

The context

Generative AI reached marketing teams through general-purpose chat tools. Teams used them for ideation anyway, and Purl started in the gap between what a chat window does and what an ideation session needs.

My role

Co-founder and founding product designer. I owned the product vision alongside the execution: research, information architecture, the design system, high-fidelity prototyping and the investor-facing material.

Team and timeline

Three co-founders, all working part time: me on product and design, an engineer building the MVP, and a third handling logistics, client relationships and the funding pitch. Six months from concept to MVP.

Constraints

No funding. The MVP existed to be pitched, so it had to look and behave like a finished product on a budget of nobody’s time but our own. I built it in evenings and weekends while working full time at LiveScore, which meant every design decision had to be cheap to build and hard to argue with, because there was no capacity for a second attempt.

Key metrics

22%

Higher task completion than working through the same problem with a conversational AI tool, measured by unmoderated testing across 22 users.

6

Months. From first concept to an investor-ready MVP, built part time alongside full-time roles.

2x

Increase in build velocity for the engineering co-founder once the Figma design system and matching component repository were in place, compared with their attempts to build it before I joined.

Problems identified

A typical AI chat interface, starting from an empty text box
The blank canvas

Users found open text inputs intimidating. They needed prompts and guardrails to begin.

Early iterations of Purl’s features
Analysis paralysis

Competing tools generated too many ideas. Users could not identify quality in the volume.

A separate review screen, cut off from where the ideas were generated
A disconnected workflow

Copying AI output into spreadsheets broke momentum and lost the context behind each idea.

Research insights

Thinking spaces

‍

People do not ideate in a linear stream. They need distinct modes for generating and for refining.

Ownership builds trust

‍

Users rejected black-box AI. They needed to see clearly what was theirs and what the model had produced in order to feel the final idea was their own.

The overlay was the moment

‍

Testing showed the deep-dive analysis panel was the highest-value feature, which validated moving away from a chat interface altogether.

Decisions & trade-offs

The central call was rejecting the chat interface most AI tools had settled on. Chat is cheap to build and familiar to users, but testing showed it was the cause of the problem, so I replaced it with a structured board.

‍

With no funding and three part-time founders, I cut roughly 40% of proposed features to protect a single path through the MVP. The hardest to lose was multi-user collaboration, which was too complex to build at that stage. Cutting it was the decision that made the six-month timeline possible.

‍

I also made the high-fidelity Figma prototype the main asset for investors, so design decisions had to carry a commercial argument as well as a usability one.

Experimentation

Purl’s Kanban board, used to test how much the AI should do on its own
Human in the loop

Tested several levels of AI autonomy. Users consistently preferred the model as a partner offering suggestions rather than a generator supplying answers.

Features mapped to each stage of the workflow to decide what to cut
Feature culling

Removed complex parameter settings and task completion improved, which made the case for fewer options.

An idea being refined on the board
Parallel versus linear

Unmoderated testing put the board against a chat interface on the same ideation task, to test whether holding several ideas open at once beat working through them one message at a time.

The solution

Purl’s Kanban board, with columns for Ideate, Refine and Review
The guided Kanban

Replaced the open chat interface with a structured board that moves users through distinct thinking spaces: Ideate, Refine, Review. Ideas persist as objects rather than messages, so several can be developed at once and nothing is lost to scroll.

Text overlays that focus the user on one action at a time
Contextual overlays

AI analysis happens in side panels, keeping the main workspace clean and the user’s own work visually distinct from the model’s.

The final review, bringing ideas and analysis together into a plan
The review engine

A final synthesis stage that turns loose ideas into a critiqued, actionable plan, answering the “what now” problem common to AI workflows.

Modular components built to match React libraries and speed up the build
A React-ready system

Built with component parity to React libraries from the start, so engineering handoff needed no translation layer.

The Idea Card adapting at each stage of the workflow
Modular cards

The core Idea Card was designed to be content-agnostic, able to hold text, images or data as the underlying models change.

The first round of ideas generated from a user’s prompt
Workflow architecture

The Kanban framework supports unlimited columns, so the product can extend into other verticals such as product management or HR without a UI redesign.

A selection of Purl’s final designs

Lessons learnt

Systems save startups

‍

Building a design system for an MVP felt like a luxury at the time. It paid for itself as soon as development started, because nothing had to be interpreted or re-specified.

Learning to reject good ideas

‍

Multi-user collaboration was a good feature and cutting it was still the right call. Turning down bad ideas is easy. Protecting a core proposition from good ones is the part that decides whether anything ships.

Part time has a ceiling

‍

We built a working MVP and a real product argument, but three founders working evenings and weekends could not generate the traction the funding conversation needed, and eventually the others chose not to continue. The design work stands up. The operating model did not, and I would now test whether a team can actually resource something before designing as though it can.

Next project