Summary
AI tools like ChatGPT have exploded in popularity with good reason: they allow users to draft, summarize, and edit content with unprecedented speed. While these generic tools can generate any type of content or perform any type of content task, the user needs to craft an effective prompt to get high-quality output, and often needs to exchange multiple messages with additional guidance and requirements in order to improve results. When you’re building an AI-powered text generation feature, such as a product description or email writer, you typically can’t expect users to craft their own prompts. And unless you’re building a chat interface, you’re unlikely to offer the ability to iteratively improve the output. Instead, your feature needs a robust prompt skeleton that combines with user input to produce high-quality output in a single response. For the designer, this means building an interface that helps users provide the exact information that creates a successful prompt. This process is more complex than simple form design or a mad-lib prompt completion tool. The user input, often including free form text fields, might be required to fill in prompt variables, but it also could change the prompt structure itself, or even override base instructions. The effectiveness of the user input significantly influences the quality of the output, underscoring the need for designers to be deeply familiar with the backend prompt architecture so they can design the frontend. Drawing on recent text generation projects, I'll demonstrate how the interface design can respond to and evolve with the prompt architecture. I’ll talk about how to determine which prompt components to make invisible to the user, which to provide as predefined options, and which should be authored by the user in free-form text fields. Takeaways How prompt structure can impact user interface design and conversely, how design can impact prompt structure Techniques to provide effective user guidance within AI generation contexts to ensure consistently high-quality output Real-world examples and learnings from recent generative AI projects in an e-commerce software product
Key Insights
-
•
Designing AI text generation systems involves unique challenges compared to general chat interfaces.
-
•
Predefined prompts can lead to higher quality outputs as they guide user input and LLM behavior effectively.
-
•
The designer must balance fixed prompts with configurable user inputs.
-
•
Tone differentiation is crucial in preventing generic content across similar product descriptions.
-
•
User testing revealed that merchants struggled to articulate their brand's tone, highlighting the need for predefined options.
-
•
Names and descriptions of tones must be distinct to avoid confusion among users.
-
•
Providing rich, detailed tone descriptions improves the effectiveness of a simple dropdown selection.
-
•
Properly structuring prompts is essential for AI accuracy and relevance in text generation.
-
•
AI tools must cater to both seasoned and novice users, making design simplicity vital for new users.
-
•
Understanding the backend structure of prompts enhances user experience and output quality.
Notable Quotes
"Building interfaces for AI features is a new and exciting challenge for product designers."
"You need to know what to ask for, how to ask for it, what information to include, what you want the output to look like."
"When you're building a specific AI feature, you don't want to make your user write their own prompt."
"Providing tone differentiated suggestions is key to ensuring merchants can differentiate themselves."
"During our user testing, we found that articulating their brand voice was much harder than expected for most people."
"We had to come up with our own list of tones, as the LLM's suggestions were inconsistent."
"Tone zones need to be distinct enough for merchants to easily spot the tone that fits their brand."
"While a single tone label can be enough, providing descriptions of tone attributes leads to even better outputs."
"The user simplifies their job by selecting a single option while benefiting from a complex set of instructions."
"Thoughtful design can simplify complex AI processes for users, leading to better outputs."
More Videos
"The power of followership is pretty amazing."
John MaedaMaking Sense of Enterprise UX
June 9, 2016
"The way to think about it is in a simple way: how do we support this thinking style?"
Indi YoungThinking styles: Mend hidden cracks in your market
January 8, 2025
"What if design teams and data teams could partner up to solve big customer problems?"
Chris ChapoData Science and Design: A Tale of Two Tribes
May 13, 2015
"We created spaces that looked like channels on a TV."
Elizabeth ChurchillExploring Cadence: You, Your Team, and Your Enterprise
June 8, 2017
"Collectives are communities of practice individuals that have an interest in a particular topic."
Rusha SopariwalaRemote, Together: Craft and Collaboration Across Disciplines, Borders, Time Zones, and a Design Org of 170+
June 9, 2022
"I believe having an ethical framework can help limit the harm typically done in design."
George AyeThat Quiet Little Voice: When Design and Ethics Collide
November 16, 2022
"As leaders, we need to cultivate our own resilience so that we can nurture our teams."
Ariba JahanTeam Resiliency Through a Pandemic
January 8, 2024
"Navigating the organization takes time and effort; understanding stakeholder dynamics is crucial."
Julie Gitlin Esther RaiceDesign as an Agent of Digital Transformation at JPMC
June 9, 2021
"Data per se are not the problem; it's the way we are using the data that can be problematic."
Matteo GrattonCan Data and Ethics Live Together?
October 1, 2021
