AI Design Collaborator & Prompt Library

Overview

As AI tools became increasingly embedded in design workflows, I wanted to move beyond ad hoc prompting and build something more intentional. I developed a structured prompting framework, RTCCF, and used it to create a custom ChatGPT collaborator and a comprehensive prompt library that helps designers get consistently useful outputs from AI across research, design, writing, and ideation tasks.

Problem

Most designers using AI tools were getting inconsistent results, not because the tools weren't capable, but because the prompts were vague or missing key context. There was no shared language or structure for how to ask AI for design-related help, which meant outputs varied wildly and a lot of time was spent iterating on bad first responses. For a design team working at pace, that friction adds up fast.

Research

I explored how other disciplines - engineering, content strategy, data science, were approaching prompt structure, and looked at where design-specific prompting was falling short. The common thread was that the best results came from prompts that were specific about role, intent, and constraints upfront.

I also observed patterns in my own prompting habits and those of colleagues to identify where clarity was most often missing.

Solution

I developed the RTCCF Framework, a five-part structure for writing AI prompts that consistently produce useful, actionable outputs:

Built on this foundation, I created two things: a custom ChatGPT collaborator configured to guide users through the RTCCF structure before producing any output, and a prompt library covering the most common design tasks - research, design systems, ideation, user flows, writing, and micro interactions.

Design Process

I started by identifying the most frequent design tasks that could benefit from AI assistance and grouped them into categories. For each category I wrote and tested multiple RTCCF prompts, refining them based on output quality until I had a reliable set that produced consistently strong results.

The custom ChatGPT was configured with a system prompt that keeps it focused, it asks clarifying questions mapped to missing RTCCF fields rather than making assumptions and running with them. Conversation starters were added to lower the barrier to entry for new users, covering the most common entry points like drafting usability tests, turning notes into PRDs, and structuring user flows.

The prompt library was documented in a structured Miro board, organised by design phase and task type, making it easy to browse and reuse without needing to rebuild prompts from scratch each time.

Final Design

The deliverables across this project were:

• RTCCF Framework A clear, teachable structure for AI prompting that any designer can pick up and apply immediately, with annotated examples across four real-world design scenarios.
• Custom ChatGPT Collaborator A configured AI teammate that follows the RTCCF workflow, asks the right clarifying questions, and keeps outputs scoped and relevant. Built for product design tasks across research, UX, and engineering collaboration.

Prompt Library

A categorised library of ready-to-use RTCCF prompts organised across seven core design disciplines:

• Research Competitive analysis, user interviews, survey design and usability planning.
• Design Systems Component documentation, design token naming and file structure.
• Design Onboarding flows, dashboard design and prompt documentation.
• Writing Product marketing copy and design-specific writing tasks.
• Synthesising Insights Turning research findings into themes, opportunities and narratives.
• Ideation Generating and structuring ideas quickly with AI as a thinking partner.
• User Flows Mapping and proposing flows for key product journeys.

Outcomes

The most significant measurable impact was an 80% reduction in time spent producing key design documentation, tasks like drafting usability test scripts, writing PRDs, and structuring user flows that previously required significant manual effort could now be completed in a fraction of the time using the custom ChatGPT collaborator and prompt library.

Beyond speed, the quality and consistency of outputs improved, with a strong structured base rather than a blank page. The RTCCF framework also had a secondary benefit of making design thinking more explicit, encouraging clearer problem framing before jumping into execution.