RIZEN Framework: Prompt engineering

We’re constantly seeking ways to improve our efficiency and unlock new possibilities. Entering the world of Large Language Models (LLMs) — such as ChatGPT or Gemini, revolutionized our design processes. But how do we effectively communicate with these Generative AIs to get the most out of them?

The answer lies in the RIZEN Framework, a structured approach to crafting prompts that guide LLMs toward creating exceptional UX deliverables. Let’s dive deeper into each step of RIZEN.

R — Role

Imagine your LLM as a skilled collaborator — a UX researcher, UI copywriter, or even a usability testing facilitator. By defining the role you need, you set the stage for the LLM to assume the appropriate expertise and produce relevant outputs.

For example, if you’re designing a new onboarding process, you might prompt your LLM: “Act as a UX researcher and conduct an in-depth analysis of user pain points during the current onboarding experience.” This role sets the tone for the LLM to focus on gathering user insights and identifying areas for improvement.

I — Instructions

Think of this step as writing a clear and concise design brief. Provide specific instructions on the desired deliverable, just like you would for any other team member.

For instance, instead of a vague request like “Help me design a better landing page,” instruct your LLM: “Generate three landing page concept variations for our new productivity app, targeting busy professionals aged 25–35.” This ensures the LLM understands the specific goal, target audience, and desired number of outputs.

S — Steps

Break down the task into actionable steps, providing the LLM with a roadmap to follow. Think of it as guiding your assistant step-by-step through the design process.

For instance, when building a persona, your prompt could be: “Create a detailed user persona for our primary user group, tech-savvy professionals aged 25–35. Include information on their demographics, motivations, frustrations, and goals.” This helps the LLM structure the persona by providing specific details to focus on.

E — End Goal

Clearly define the purpose and target audience of the LLM’s output. This ensures the generated content aligns perfectly with your project’s vision.

For example, while designing a user flow for a mobile app, you might state: “This user flow should cater to first-time users and guide them through the app’s core functionalities seamlessly and intuitively.” This sets a clear goal for the user flow and aligns the LLM’s efforts with the desired user experience.

N — Narrowing

Just like any design project, provide constraints and specific guidelines to refine the LLM’s output. This helps ensure the results are consistent and meet your design standards.

For instance, when creating a user interface mockup, you might specify: “Use a minimalist design aesthetic with high-contrast colors. Maintain consistency with our existing design system and prioritize a mobile-first approach.” These guidelines ensure the LLM generates a UI that aligns with your overall brand and design language.

Let’s go through it ones again

R — Role

• Define the AI’s role based on the task at hand. Examples:
• “Act as a UX researcher…”
• “Act as a UI copywriter…”
• “Act as a usability testing facilitator…”

I — Instructions

• Give the AI clear instructions on the desired deliverable. Think of it like writing a design brief:
• “Create a user flow for our new onboarding process…”

S — Steps

• Outline the process you’d follow as a UX designer. Provide a structured pathway for the AI:
• “Start by identifying key user pain points, then map out decision points and potential interactions. Include alternative paths for error states.”

E — End Goal

• State the purpose and target audience of the deliverable. This aligns the AI’s output with your project’s vision:
• “This persona should represent our primary user segment: tech-savvy professionals aged 25–35 interested in productivity tools.”

N — Narrowing

• Set constraints or style guides. This helps refine the AI’s output:
• “Limit the user flow to a maximum of 5 steps.”
• “Use friendly and conversational language for the UI copy.”
• “Adhere to our existing design system for visual elements.”

Example

Designing a Landing Page

Let’s say you want a compelling landing page concept. Your RIZEN prompt might look like this:

• Role: Act as a UI/UX designer specializing in landing pages.
• Instructions: Generate a landing page concept for our new time-tracking app.
• Steps: Start with a bold headline and value proposition. Include a clear call to action, social proof, and visuals aligning with our brand.
• End Goal: Target busy professionals looking to streamline their workdays.
• Narrowing: Use a minimalist aesthetic. Prioritize mobile-first design.

Benefits of RIZEN for UX

• Rapid Ideation: Generate multiple concepts or variations quickly.
• Streamlined Persona Development: Get rich data points for building user personas.
• Efficient Copywriting: Create user-centric copy for interfaces and marketing materials.
• Scalable User Testing: Design AI-powered usability test scenarios.

Remember: The RIZEN framework is customizable! Adapt it to various UX tasks and experiment to discover the best ways to collaborate with your AI design assistant. Experiment, adapt, and explore different prompt variations to see what works best for your specific needs and the capabilities of your chosen LLM.

Ready to take your design process to the next level? Start incorporating the RIZEN framework into your workflow and witness the transformative power of AI in UX.

Want to explore more ways to leverage AI for a smoother, smarter design process? Follow me on LinkedIn for additional tips and tricks: https://am.linkedin.com/in/khatchadour

Let’s unlock the full potential of AI and revolutionize the way we design!