Chapter 5 Content Design for LLM-Based Experiences
[ 5 ]
Content Design for LLM-Based Experiences
Responsibility lies with humans at the end of the day.
—Dr. Joy Buolamwini, President of Algorithmic Justice League
The majority of this book describes static UX content: the descriptions, titles, buttons, and more that help people understand how to use the experience. Dynamic content, in contrast, is content generated on the fly.
Content can be tuned to maximize its effect. That is, when an experience relies on content to convey the value, purpose, or actions available, different content will have different effects on the success of that experience. It’s worth investing human effort in the static content that every person will see the same way to make sure it’s effective in all its circumstances. But dynamic content, intended to be unique hundreds to millions of times each day, can be handled by large language models (LLMs).
LLMs are well-suited1 for creating large volumes of content that is repetitive in its purpose, style, and format. For example, consider an LLM-based feature that generates product descriptions for a digital marketplace. Each product in the marketplace may differ wildly from the next, but each of them needs a title and description that includes recognizable keywords and attributes.
For another example, consider the questions that customers ask support specialists in call centers. Each person’s specific circumstances may be slightly different, but most of the ...
Become an O’Reilly member and get unlimited access to this title plus top books and audiobooks from O’Reilly and nearly 200 top publishers, thousands of courses curated by job role, 150+ live events each month,
and much more.
Read now
Unlock full access